How to use Reverbot

Before you begin

• An iPhone 12 or newer with iOS 26 or later. On an older iPhone, Reverbot shows “Reverbot needs a newer iPhone”.
• You must be 18 or older. The AI services Reverbot uses require it.
• An internet connection. Answers, translations and most voices come from online services.
• Headphones, to keep answers private. Without them, answers appear on screen instead of playing aloud, unless you turn on Speakerphone in Settings (the gear at the right end of the top bar).
• Access to the microphone and to speech recognition. Reverbot can’t hear you without both.

You don’t need an account, a card or an AI key to start.

Set up Reverbot the first time

1. Open Reverbot. It makes a guest account for you, so there’s nothing to sign up for.
2. On “Let Reverbot hear you”, tap Allow microphone.
3. When iPhone asks, allow access to the microphone and to speech recognition.
4. Optional: On “Your profile”, enter your name in Name people see and choose a username. If you don’t choose one, Reverbot picks a username for you.
5. Tap Continue.
6. Read the short tour, tapping Next on each page, or tap Skip tour.
7. Tap Start using Reverbot.

A walkthrough then starts on the main screen. It points at the big button and asks you to try an example question; its answer is a prepared example, so it doesn’t use your free questions.

If you tap Skip on “Let Reverbot hear you”, iPhone asks for access the first time you start listening.

Note: If Reverbot can’t make a guest account, for example with no internet on first launch, “Welcome to Reverbot” appears. Check your connection and tap Start without an account, or sign up there.

Agree to the AI data notice

The first time you tap the big button, the “Before we start” screen tells you what leaves your iPhone: which AI companies receive what you say, that documents you add go with your questions, which voice services receive the words Reverbot speaks and when anything is recorded.

1. Read the screen. To read everything, tap Privacy policy.
2. Turn on I’m 18 or older.
3. Tap Agree and continue.
4. Tap the big button again to start listening.

If you tap Not now, Reverbot can’t answer anything, and the screen appears again the next time you start listening. It also appears again if the notice changes.

Ask your first question

1. On the main screen, tap Questions under Listening mode.
2. Tap the big button.
3. Wait until “Listening for questions...” appears below the big button.
4. Ask a complete question, such as “What is the capital of Japan?”
5. Stop talking and wait. Reverbot answers after a short pause.

Your words appear under the big button as you speak. The status line above it changes from “Listening” to “Processing” while the answer is prepared, then to “Speaking” while it plays.

Hear or read the answer

With headphones connected, the answer plays in them and appears under the big button as it’s spoken.

Without headphones, Reverbot shows the answer instead of playing it aloud, with “Connect headphones to hear it” above it. To hear it, connect headphones: the answer on screen plays in them. Or tap Speakerphone beside the message.

Important: Tapping Speakerphone there also turns on Speakerphone in Settings, so every later answer plays out loud too. To keep answers private again, open Settings and turn off Speakerphone.

To change how much detail answers have, tap Concise under Input & answer to switch to Extended, and tap it again to switch back.

Continue, repeat or stop

• After an answer, Reverbot keeps listening. Ask your next question when “Listening for questions...” appears again.
• To hear the last answer again, tap Repeat that under the big button.
• To stop an answer while it plays, tap Tap to skip. Reverbot keeps listening.
• To stop listening, tap the big button. The status line reads “Inactive”.

Note: Choosing another pill or opening Settings doesn’t stop listening. Only the big button does.

Know your free questions

Free questions are answered by Reverbot’s AI at no cost to you. They’re a one-time grant and never refill:

• Before you subscribe, you have 5 free questions.
• Your first subscription, including its free trial, gives you 50 free questions in total.

The count shows above the big button, for example “3 questions left”, and reads “Questions used” when none are left. Before your trial, Start your 3-day free trial sits beside it. Once your trial has started, Get your key — required later takes its place, and reads Get your key — required now when your questions are used.

When nothing can answer, the line under the big button says what’s needed: “Add a free AI key to start asking”, “Start your free trial to keep asking” or “Subscribe to keep asking”. Tapping the big button then opens that step.

Set up a free Gemini key

Your own AI key keeps answers coming after your free questions. A Gemini key from Google AI Studio is free within Google’s daily limits, which AI Studio shows. You need a Google account to make one, even if you use Reverbot with Apple.

1. Above the big button, tap Get your key — required later, or Get your key — required now. The link appears once your trial has started.
2. On Get your key ready, or Your AI key is required now, tap Set up a free Gemini key. The Set up Gemini screen opens.
3. Tap Get my Gemini key. Google AI Studio opens in Safari; sign in if Google asks.
4. If no key is shown, tap Create API key on Google’s page.
5. Tap the copy icon next to the key.
6. Return to Reverbot, then tap Paste.

“API key saved — Gemini selected” appears, and the screen closes by itself. Reverbot now answers on your key.

Note: On Google’s free tier, Google may use what you send to improve its products, and people at Google may read it, except in the EEA, Switzerland and the UK.

Add a key in Settings

You can add a key at any time, before or during a trial.

1. Open Settings › AI & API Keys. The AI Provider & Keys screen opens.
2. Tap the sliders button (three short lines with knobs) beside Gemini.
3. Under How to get one, tap Get your Gemini key, then make and copy a key in AI Studio.
4. Return to Reverbot and paste the key into Gemini (Google) API Key. It saves as you paste.

Back on AI Provider & Keys, the summary at the top shows Gemini answering “On your own key”.

To use ChatGPT, Claude or Grok instead, tap Other AI providers under Advanced. Each needs a paid key; the API keys topic explains every key.

Create an account

You can use Reverbot as a guest. Create an account to use the web portal at reverbot.ai, recover synced settings, get replies from support and subscribe. Local-only history, such as Chat messages, has its own recovery limits. Everything you did as a guest moves to the new account.

1. Open Settings.
2. Near the end, under Account, tap Create an account or sign in.
3. Tap Sign up with Apple or Sign up with Google, or enter an email and a password and tap Create account.

To sign in to an account you already have, tap Log in at the top of the same screen. If you tap Start free trial or Subscribe as a guest, Reverbot asks you to create an account first.

Link a Google account

Reverbot works with Sign in with Apple and without a Google account. A Google account is needed for optional services such as creating a Gemini key in Google AI Studio; your own Google Cloud voice setup has separate billing and key requirements.

1. Open Settings.
2. Under Google account, tap Sign in with Google.
3. Choose the account on Google’s page.

This does not change how you sign in to Reverbot. Use a different Google account changes the linked account. Review the voice service’s current eligibility in Voice & Speed rather than assuming a linked account pays for every Google service.

Find help and the walkthroughs

• Tap the ? button at the left end of the top bar to open How to use Reverbot. Tap a topic, or search for a control or task.
• In Settings, How to Use Reverbot at the top opens the same Help.
• In a mode’s settings, How [mode] works opens Help at that mode’s topic, for example How Questions works.
• Several topics start with a row such as Walk me through Questions. It closes Help and shows each control on the real screen, free.
• To replay the first walkthrough, open Settings and, under Getting Started, tap Show the walkthrough again.

If it doesn’t work

• If tapping the big button shows “Before we start” again, turn on I’m 18 or older, tap Agree and continue, then tap the big button.
• If no words appear under the big button, open iPhone Settings, go to Apps › Reverbot, and turn on Microphone and Speech Recognition. Then tap the big button twice to restart.
• If your words appear but no answer follows, ask again as one complete question, and read any message about your connection, free questions or key.
• If the answer appears but you don’t hear it, connect headphones or tap Speakerphone, and check the volume.
• If the line under the big button reads “Add a free AI key to start asking”, your free questions are used: add a key in Settings › AI & API Keys.

Still not working? In Settings, under Feedback, tap Contact Support and say what you tapped, what you expected and what happened. As a guest, create an account first if you’d like a reply.

What the subscription includes

• 50 free questions in total with your first subscription, answered by Reverbot’s AI. Renewing doesn’t add more.
• Translation without using your free questions, from a daily allowance: Translate Vicinity, Translate 2-way, Translate with someone nearby and Translate too in an AI Meeting.
• Reverbot Voice, Reverbot’s own cloud voice, from a daily share. Without a subscription it stops when your free allowance is used, and answers continue in the iPhone’s own voice.
• Automatic emailing of AI Meeting reports to participants, when you turn it on.

Answers past your free questions need your own AI key, with or without a subscription.

Note: A subscription to ChatGPT, Claude or Grok is separate. It isn’t an API key and doesn’t pay for answers in Reverbot.

Start a free trial or subscribe

Important: A free trial is offered once per Apple Account. When it ends, Apple charges the subscription price each period until you cancel.

1. Open Settings.
2. Under Subscription, tap Start free trial, or Subscribe if no trial is offered. You can also tap Start your 3-day free trial above the big button.
3. Read the screen that opens: what you get, such as “50 questions with your first subscription”, the price, the billing period and any free trial.
4. Tap Start free trial or Subscribe.
5. If you’re using a guest account, sign up with Apple, Google or email when Reverbot asks. The purchase then continues.
6. Confirm the purchase with Apple.

Under Subscription, Reverbot then shows “Active”.

Understand your free questions

• Every account starts with 5 free questions.
• Your first subscription gives you a new one-time grant of 50 free questions in total. They never refill, and renewing doesn’t add more.
• Each answer from Reverbot’s AI uses one. An AI Meeting report can use several, and shows how many before it runs. Before you subscribe, each translated sentence uses one too.
• The count above the big button shows what’s left, for example “12 questions left”. It turns orange near the end and reads “Questions used” when none are left.
• With your own AI key, answers don’t use free questions, and the count goes away.
• If you’ve subscribed before, the subscription screen says “Your original question grant stays”: subscribing again doesn’t add more.

Note: A question Reverbot doesn’t answer, such as small talk in Questions, doesn’t use one.

Keep answers going with your key

Your own AI key answers after your free questions are used, with or without a subscription. A Gemini key is free within Google’s daily limits; ChatGPT, Claude and Grok keys are billed by their companies.

• Once your trial has started, the link beside the count above the big button reads Get your key — required later while you have questions left. Set up your key early.
• When they’re used, it reads Get your key — required now, and the line under the big button reads “Add a free AI key to start asking”.
• Tap either link, then Set up a free Gemini key, for a guided setup. Or open Settings › AI & API Keys at any time, before or after a trial.

Your key’s company sets its own limits and bills you directly. Reverbot’s subscription doesn’t pay that bill.

Understand translation costs

• Before you subscribe, each translated sentence uses one free question. Speech already in your language costs nothing.
• With a subscription, translation uses a daily allowance instead. When a day’s is used, Reverbot says “That’s all the translation for today — it comes back tomorrow.”
• A key you use for answers doesn’t pay for translation. To translate on your own account, open Settings › AI & API Keys, tap Translation under AI for each feature, and choose Gemini or ChatGPT. Google or OpenAI then bills you for each sentence.
• On a phone call with Both of you, Reverbot always uses its own translation.

With your own key and no subscription, a translation mode shows what’s left above the big button, for example “3 free questions left for translation”, with Subscribe to keep translating.

Restore a subscription

Restore your subscription after reinstalling Reverbot or moving to a new iPhone. It stays with the Reverbot account it was first used with, so sign in to that account first.

1. If you’re using a guest account, open Settings and, under Account, tap Create an account or sign in.
2. Tap Log in and sign in the way you did before.
3. Make sure iPhone uses the Apple Account that bought the subscription.
4. In Settings, under Subscription, tap Restore Purchases.
5. Read the message under the button: “Your subscription is restored.” or “No active subscription found for this Apple ID.”

If Sign in to use your subscription appears above the big button, the subscription belongs to another Reverbot account: tap it and sign in with that account. Buying again doesn’t restore anything.

Change or cancel the subscription

1. Open Settings.
2. Under Subscription, tap Manage subscription. Apple’s subscription settings open.
3. Change or cancel the subscription there.

Important: Deleting Reverbot, signing out or deleting your Reverbot account doesn’t cancel the subscription. Apple keeps billing until you cancel it with Apple.

Cancel at least 24 hours before the current period ends to avoid the next charge.

Important: Delete account, at the end of Settings, erases your settings, uploaded documents, question history and any money on your phone-calls balance. It can’t be undone.

If it doesn’t work

• If the subscription screen shows Retry instead of a price, check your internet connection and tap Retry.
• If you see “Your purchase is awaiting approval.”, wait: Reverbot updates by itself when Apple approves it.
• If Sign in to use your subscription appears above the big button, you’re signed in as a guest or to a different Reverbot account. Tap it and sign in with the account you subscribed with.
• If “Your subscription couldn’t be verified on this device.” appears, tap Restore Purchases in Settings, under Subscription.
• If translation stops for the day, wait until tomorrow, or choose your own Gemini or OpenAI key for Translation under AI for each feature.
• If answers stop after your free questions, tap Get your key — required now above the big button and set up a key.

Still not working? In Settings, under Feedback, tap Contact Support. Never include an API key or a password.

Choose a listening mode

1. On the main screen, tap a pill under Listening mode. It lights up.
2. Tap the big button to start.

Only one listening mode runs at a time. If Reverbot is already listening, the new mode takes over at once. Tapping a lit pill doesn’t stop listening; only the big button starts and stops.

• Questions answers factual questions.
• Dictionary explains words and phrases.
• AI Meeting records a meeting everyone agrees to, then writes a report.
• Translate Vicinity translates the languages around you into yours.
• Translate 2-way translates a conversation in both directions.
• Flirt suggests what to say, privately in your headphones.
• Custom Conversation coaches a conversation you set up, or answers from your own material.

Note: Job and Personal Assistant aren’t pills: they start with a wake phrase. In Settings (the gear), under Wake Phrases, turn on Job or Assistant, then say “hey job” or “hey assistant” while Reverbot is listening.

Add Questions to a translation

Questions can run beside Translate Vicinity or Translate 2-way. Reverbot translates what it hears first, then answers any question in it.

1. Tap Translate Vicinity or Translate 2-way.
2. Tap Questions. Both pills light up.
3. Tap the big button.

You can also tap Questions first, then the translation pill. The line under the big button says both are on, for example “Listening — foreign speech is translated to English, then questions are answered...”.

To stop answering, tap Questions again. If you tap the lit translation pill instead, translation turns off and Questions stays on.

Note: Each answer uses one of your free questions, or your own AI key. If your free questions run out, translation carries on.

This is the same switch as Answer questions too in the Vicinity and Two-way settings.

Open a mode’s settings

1. Touch and hold the mode’s pill, then tap Settings.
2. Change what you need. Most changes apply at once; a note under a setting says when one waits for the next start.
3. Tap Done.

You can also open Settings and, under Listening Mode, tap the sliders button (three short lines with knobs) beside a mode. There, the rows for the translation and custom pills are named Vicinity, Two-way and Custom, and their screens are titled Vicinity Language Translation, Two-way Language Translation and Custom Conversation.

Every mode’s screen starts with How [mode] works, which opens its Help topic, then Button style. It ends with Restore [mode] defaults, for example Restore Questions defaults. In between are the mode’s own settings, explained in its topic, and in most modes:

• Silence duration before answering, with How long it waits: the pause after you stop talking before Reverbot replies.
• AI: which AI the mode uses, or the default from AI & API Keys.
• Voice: Different voice for [mode] and Different speed for [mode].

Restoring defaults keeps saved conversations, documents, locations and meeting history.

Use the Custom Conversation pill

The Custom Conversation pill runs the scenario or information source you’ve activated. With none activated, tapping it opens its settings instead.

1. Tap Custom Conversation, or touch and hold it and tap Settings.
2. Tap Your scenarios and sources. The Custom Conversations screen opens.
3. Tap the circle beside a template to activate it, or tap the Add button (+) to make one.
4. Go back and tap Done.
5. Tap the pill to choose it. It shows the template’s Main-screen button name, up to 15 characters, or Custom Conversation if it has none.

To deactivate a template, tap its circle again. The Custom Conversation topic explains how to write a scenario.

Until a template is active, the Custom row in Settings, under Listening Mode, reads “Make or choose a scenario first — tap here.”

Use the Input & answer pills

The pills under Input & answer change how Reverbot listens and answers:

• Room or Solo: tap to switch the microphone. Room hears everyone around you; Solo isolates your voice and reduces background sound.
• Concise or Extended: tap to switch how much detail answers have. It applies from the next answer and isn’t shown in Flirt.
• Speaker or Earpiece: shown in Translate 2-way when Phone speaks your reply is on, and in an AI Meeting translating out loud. Speaker plays your translated words from the loudspeaker; Earpiece plays them quietly, for someone holding the phone to their ear.

Flirt, both translations, Custom Conversation, AI Meeting and every rehearsal set the microphone themselves, so the microphone pill can’t be tapped while one is chosen. Your own choice comes back when you leave the mode.

The same choices are in Settings, under Microphone Input and Answer Style. The Input & answer topic explains them in detail.

Turn on Voice Changer

Voice Changer changes how you sound to the people you talk to in calls made through Reverbot. Its effects run on your iPhone, are free and work alongside any listening mode.

1. Tap Voice Changer under Listening mode. The Voice Changer screen opens. It’s also in Settings.
2. Turn on Voice Changer. Its pill lights up.
3. Under Free voice effects, tap Voice and choose an effect, such as Deep voice, Robot or Telephone.

The line under the switch shows the effect in use, for example “On · Robot”. Under Where to use it, choose where your changed voice goes:

• Phone & Reverbot calls: calls to phone numbers and Reverbot calls. On by default.
• Video & group calls: video calls and meetings in Chat. On by default.
• AI spoken replies: Reverbot’s own spoken answers and a live practice partner’s voice. Off by default.

Note: Calls in Apple’s Phone app, FaceTime and other apps can’t use Voice Changer. Speech recognition always hears your own voice, and call permission announcements keep their normal voice.

Fine-tune and save an effect

1. On the Voice Changer screen, under Free voice effects, tap Fine-tune your voice.
2. Drag the sliders you want: Pitch, Vocal character / formant, Bass, Treble, Robot, Robot frequency, Distortion, Echo, Room, Breathiness, Voice tremble and Output level.
3. Optional: Choose a Digital texture for a lo-fi sound.
4. Enter a name in Name your effect.
5. Tap Save custom effect.

Your effects appear under the effect list. Tap one to use it, or swipe left on it to delete it. You can keep up to 30, with names up to 40 characters.

Tip: Change one slider at a time, and listen with the microphone studio and headphones before using the effect on a call.

Try an effect or record

The microphone studio, on the Voice Changer screen, lets you hear and record your changed voice. It uses the microphone on its own, so end any call or listening session first.

1. Under Microphone studio, tap Start microphone preview.
2. With headphones connected, turn on Hear my changed voice to hear yourself live.
3. Tap Record changed voice and speak.
4. Tap Finish recording. Recording stops by itself at 5 minutes.
5. Tap Play recording, Save or share recording or Delete recording.
6. Tap Stop microphone preview when you’re done.

To start from a file, tap Import an audio recording and choose one up to 5 minutes and 40 MB.

Under Free soundboard, tap Chime, Laser, Robot beeps, Celebration, Censor beep, Drum hit, Applause or Magic to play it into the studio or into a call made through Reverbot. Voice Changer must be on, and muting your microphone mutes the sounds too.

Use Voice Changer on a call

Every call screen in Reverbot, for phone calls, Reverbot calls and video calls, has a Voice Changer button. It reads Voice Changer · On while your voice is changed.

1. During the call, tap Voice Changer.
2. Turn Voice Changer on or off, choose a Free voice effect or drag Pitch.
3. Optional: Under Saved effects, tap one of your effects. Under Soundboard, tap a sound.
4. Tap Done.

Your call stays connected while you change anything, and your changes are saved for next time. Apply to phone & Reverbot calls and Apply to video calls are the same switches as Where to use it on the Voice Changer screen. Fine-tuning and recording are only on that screen.

Convert a recording with AI

Important: Conversion uploads your recording to ElevenLabs, Respeecher or Resemble and is paid by your own account there. Reverbot shows what will be sent before anything is uploaded, and never retries by itself.

1. Record or import a recording in the microphone studio.
2. Under Natural AI voices · your own account, choose ElevenLabs, Respeecher or Resemble.
3. Paste your key for that company. For Respeecher or Resemble, tap Save key. How to get a [company] key shows where to make one.
4. Tap Load my [company] voices.
5. Choose a voice in Target voice.
6. Tap Convert recording with [company].
7. Read the message, then tap Upload & convert.

When it’s ready, tap Play converted voice, Save or share converted voice or Delete conversion. Respeecher can take a few minutes.

Note: Resemble needs your own Cloudflare storage, set up from Video calls & recordings on the same screen. Your recording is deleted from it as soon as Resemble has it.

Use a live AI voice

A live AI voice on calls needs Resemble Live VC, with a server and voice model your provider sets up for you. An ordinary Resemble API key isn’t enough.

Important: While it runs, your microphone audio goes to your provider, and your provider charges your own account. Network and Bluetooth add delay.

1. On the Voice Changer screen, under Advanced live AI conversion, tap Resemble Live VC · provisioned accounts.
2. Enter Provider-issued HTTPS hostname, Live conversion API key and Provisioned live voice / model, then tap Save live API key.
3. Turn on My provider agreement permits this app and voice use.
4. Set Session limit, from 1 to 30 minutes.
5. Start a call or the microphone studio, then choose it in Active audio route.
6. Tap Start paid live conversion, then tap Start live conversion & provider usage to confirm.

If the connection drops, your changed voice stays muted until you tap Stop AI conversion & use local effect. It never reconnects or charges again by itself. Once set up, a call’s Voice Changer button also offers Start live AI conversion.

If it doesn’t work

• If nothing happens after you tap a pill, check that it lit up, then tap the big button to start.
• If the Room pill can’t be tapped, the chosen mode or a rehearsal needs to hear the people around you. Choose Questions or Dictionary to use Solo.
• If Custom Conversation opens its settings instead of lighting up, activate a template first in Your scenarios and sources.
• If Concise is missing, Flirt is chosen. Flirt doesn’t use answer length.
• If people on a call don’t hear your changed voice, check that Voice Changer is on, and that Phone & Reverbot calls or Video & group calls is on. Calls in Apple’s Phone app and FaceTime can’t be changed.
• If you see “Finish the current call, listening session or recording first.”, stop listening or end the call, then tap Start microphone preview again.
• If the soundboard is dimmed, turn on Voice Changer and start a call or the microphone preview.

Still not working? In Settings, under Feedback, tap Report a Bug and describe the steps, what you expected and what happened.

Adjust one feature

1. Touch and hold the feature’s pill, then tap Settings.
2. Find Silence duration before answering.
3. Move the slider toward Shortest or Longest.
4. Read the timing explanation below it, then try a complete sentence.

For all features in one place, open Settings › Silence duration before answering.

Choose a useful pause

Start at Natural, the fixed default marked in the middle. If Reverbot cuts in while you are still speaking, move toward Longest. If it waits longer than you prefer after complete sentences, move toward Shortest a little at a time. Test with your normal pauses for thought. Changes update the running feature’s timing.

Understand the displayed time

The displayed wait includes the feature’s speech-recognition or turn-detection delay. Shortest is each slider’s minimum; translation can go down to 0.1 seconds. This controls when processing begins, not the total time until an audible reply: the request and voice preparation take additional time. Read the explanation under the slider for the selected feature.

Understand how a turn ends

Questions, Dictionary, Job, Personal Assistant, Home Control, Souls, Flirt, Custom Conversation and Practice can use a completion check. At Natural or faster, an available completion check can let them respond before the selected pause expires. Translate Vicinity, Translate 2-way and Life Journal use pause-based timing. Live rehearsals can use their voice partner’s own turn detection; the Practice slider explains the selected partner’s behavior.

Avoid splitting sentences

Very short settings can treat a breath as the end of a turn. Read the orange warning if it appears. Features that assess the words allow one second more for an unfinished sentence. Vicinity and Two-way translation do not add that extra second: they end captured segments from silence, so making them too fast can translate fragments separately.

Change the setting used by an extra

Answer questions too uses the Questions timing. Translate too uses Vicinity translation timing. Translate nearby uses Two-way translation timing. Adjust those named sliders when an added feature feels too quick or slow; changing the main mode’s slider does not change every extra attached to it.

Restore the defaults

Tap Back to default beneath a changed slider to restore that feature. To reset every feature’s pause, open Settings › Silence duration before answering and use Restore all defaults, then confirm. These controls reset response pauses, not your saved conversations or content. If a long delay remains at the default, check the connection and any provider message instead of making the pause increasingly shorter.

If it doesn’t work

• If Reverbot interrupts mid-sentence, open that feature’s Settings and move How long it waits toward Longest.
• If the pause feels too long, start at Natural and move toward Shortest a little at a time.
• If “Processing” remains after you finish speaking, the AI or voice service is still preparing a response. Changing the silence setting does not repair a connection or provider limit.
• If only an extra answer or translation feels slow, change the corresponding Questions or translation slider.

Still not working? In Settings, under Feedback, tap Report a Bug and include the feature and timing setting.

Start listening

1. Tap a pill under Listening mode.
2. Tap the big button.
3. Wait for the line under the big button, such as “Listening for questions...”, then talk.

Note: The first time, “Before we start” appears instead. Agree to it and allow the microphone and speech recognition, then tap the big button again.

While idle, the big button is dim with a microphone, and the line under it says what a tap does, such as “Tap the button to start listening”. In Flirt it’s pink with a heart. Three rings pulse around it while a session runs.

Important: When nothing can answer, a tap opens what’s needed instead: your own AI key, a free trial or a subscription. The line under the big button says which, for example “Add a free AI key to start asking”.

Tip: With VoiceOver on, a two-finger double-tap anywhere starts or stops listening.

Read the status line

The status line above the big button, with a colored dot, shows what Reverbot is doing:

• “Inactive”, gray: nothing is running.
• “Listening”, cyan: waiting for speech. Your words appear under the big button as you speak.
• “Processing”, purple: preparing an answer. Your question stays on screen.
• “Speaking”, green: the answer plays and appears under the big button.

In Flirt, it reads “Listening to them”, “Thinking...”, “Speaking suggestion” or “Listening to you”.

During a wake-phrase command, it reads “Listening for commands”, “Waiting for details...”, “Working...” or “Done. Say undo to revert.”

The line under the big button names each mode’s own step, such as “Listening for words...” in Dictionary or “Recording the meeting...” in AI Meeting.

Skip, repeat or hear an answer

• While an answer plays, tap Tap to skip under the big button. Reverbot stops that answer and keeps listening.
• After an answer, tap Repeat that to hear it again.
• If “Connect headphones to hear it” appears, Reverbot is showing the answer instead of playing it, because no headphones are connected and Speakerphone is off. Connect headphones and it plays in them, or tap Speakerphone.

Important: Tapping Speakerphone there also turns on Speakerphone in Settings, so later answers play out loud too. Turn it off in Settings to keep answers private.

Headphone gestures and shaking iPhone can skip answers too; the Hands-free controls topic explains how.

Stop listening

Tap the big button again. The status line reads “Inactive” and the microphone turns off.

• In Flirt, a conversation card takes the big button’s place while it runs. Tap End Conversation.
• During a call with a soul, the big button shows the soul. Tap it to hang up.
• Locking iPhone doesn’t stop a session. Reverbot keeps listening in your pocket.
• Choosing another pill switches modes without stopping.

Note: In Flirt and in Custom Conversation coaching, Reverbot stops by itself if your headphones disconnect, so nothing private plays out loud. “Headphones disconnected” appears: reconnect them and tap the big button to start again.

Choose your default button style

The button style is the animation in the big button while Reverbot listens or talks. Your default style is used by every mode without a style of its own.

1. Open Settings (the gear), then tap Center button appearance. The Button style screen opens.
2. Under Your default style, tap a style.
3. If the style opens a screen of its own, choose a color there, or set up the face and tap Use [style], for example Use 3D face.

The top of the screen shows the style in use; the previews in the list move with sample speech. Button style at the top of Settings › Voice & Speed opens the same screen.

Note: A saved Custom Conversation can have its own style, set in its editor, and a soul has its own, set in its look.

Give one mode its own style

From the main screen, for the mode you’re in:

1. Tap the sliders button (three short lines with knobs) under the idle big button. Button & Voice opens on Button style.
2. Turn on Only for [mode], for example Only for Questions.
3. Tap a style.
4. Tap Done.

With Only for [mode] off, a style you tap here changes your default instead.

You can also open Settings › Center button appearance, tap a mode under Each mode (Life Journal is there too) and tap a style. Or, in a mode’s settings, under Button style, turn on Different button style for [mode] and tap Style. To put every mode back on your default, tap Reset all mode styles to default on the Button style screen.

The Voice part of Button & Voice sets the voice and speed for the same mode.

Turning Only for [mode] on after a style choice keeps the choice for that mode and restores the default this page changed. In a saved scenario, the style switch names that scenario; its voice switch applies to Custom Conversation.

Pick a style and color

• Light and glow: Reverbot blue (the default), Soul glow, Luminous orb, Glass Light - Matte, Glass Light - Shiny and Ion sphere.
• Sound pictures: Wave, Voice memo, Retro equalizer, Oscilloscope and Talking car.
• Talking faces, free on your iPhone: Cartoon face, Cartoon lips, 3D face and 3D lips.
• A consenting person’s recorded face: Personal face and Personal lips.
• Off: a still waveform.

Luminous orb, Ion sphere and both Glass Lights come in five colors. Tap the style, then under Choose a color, tap Original, Red, Orange, Purple or Pink.

Each style except Off moves with the sound of the voice. With Reduce Motion on in iPhone Settings, the display stays still.

Design a cartoon character

1. In the style list, tap Cartoon face.
2. Tap Character studio.
3. Tap a starting character, such as Scholar or Stargazer, or tap Surprise me. Undo appearance change goes back one step.
4. Under Face & hair and Details, change what you like, such as Hair, Eyes, Glasses or Freckles.
5. Under Decorative frame, choose an Accent color and a frame.
6. Go back, then tap Use Cartoon face. It changes to “Selected”.

Frames sit inside the circle and glow with speech; they contain no buttons. Cartoon lips draws a mouth only: its studio has Lips, with Lip color and Mouth size, and Surprise me instead of starting characters.

Use a free 3D face

1. In the style list, tap 3D face, or 3D lips for a close-up of the mouth.
2. Under Appearance, choose a Face. There are eight, such as Alex, Jules and Sage.
3. Optional: Tap Character studio to change Face width, and choose an Accent and a Decorative frame.
4. Tap Use 3D face.

3D faces speak the words, blink and show a soul’s mood. Each has its own hairstyle and features; to design every detail, choose Cartoon face instead.

Note: The mouth matches the words most closely in English and less precisely in other languages. Speech can take a moment to start.

Record a personal face

Important: Record only someone who is there with you and agrees. They say a consent statement on camera, and their face is shown labeled “AI portrayal”.

1. In the style list, tap Personal face, then tap Record a consenting person’s face.
2. Enter Person’s name and turn on The person agrees to this recording and AI portrayal.
3. Tap Record face & permission, and allow the camera and microphone if iPhone asks.
4. Have the person say their name and the statement on screen, then look forward, turn slowly from side to side and speak. Record 3–30 seconds.
5. Drag the slider to a clear, forward-facing frame.
6. Turn on The person gave the permission above on this recording.
7. Optional: Turn on Permission to share this AI face or Permission for use after their death, only if the person said so on camera.
8. Tap Save consented face, then tap Use Personal face.

The face is saved encrypted on this iPhone and talks for free: its mouth and jaw move with the words. Nothing is uploaded.

Make a personal face realistic

A personal face can also talk with a realistic animation paid on your own account. Until one is set up, the free animation is used.

Important: Each option uploads the chosen portrait to that company and charges your own account. Reverbot shows the cost and terms before anything is uploaded.

1. In Personal face, tap How [name]’s face talks, for example How Mia’s face talks.
2. Choose Live realistic face · Anam or Realistic clips · Simli. Free, on this phone goes back to the free animation.
3. Paste your key in Your Anam API key or Your Simli API key.
4. Tap Save Anam key or Save Simli key.
5. Turn on that company’s permission switch.
6. Tap Upload portrait & make live face for Anam, or Upload portrait & create paid face for Simli, then confirm.

Anam streams a live face while it’s connected, billed by the second, up to 3 minutes at a time. Simli makes a clip for each sentence, and each sentence waits for its clip; tap Check readiness until the face is ready. Remove local face & recording deletes the face from this iPhone, and can delete the Anam face too; a Simli copy stays in your Simli account.

Adjust button vibration

Button vibration is shared across the app and starts on. It appears under Center button in Settings, and on Button style in Button & Voice.

1. Open either location and leave Button vibration on to feel the effect.
2. Drag Strength below the midpoint for less vibration or above it for more. The midpoint is the default.
3. Release the slider to feel a sample.

Turn Button vibration off to disable it everywhere.

Note: The Only for [mode] switches scope style and voice choices. They do not make vibration mode-specific.

If it doesn’t work

• If a tap on the big button opens another screen instead of listening, read the line under the big button: it says what’s needed first.
• If the status says “Listening” but no words appear, open iPhone Settings, go to Apps › Reverbot, and turn on Microphone and Speech Recognition.
• If an answer appears but you don’t hear it, connect headphones or tap Speakerphone.
• If one mode shows a different style from the others, it has its own: open Settings › Center button appearance and check that mode under Each mode.
• If Use Personal face is dimmed, record a consented face first, then choose it.
• If a realistic face doesn’t appear, the free animation runs until Anam or Simli is set up. Open How [name]’s face talks and read the note there.

Still not working? In Settings, under Feedback, tap Report a Bug and describe the steps, what you expected and what happened.

Before you begin

• When iPhone asks, allow access to the microphone and to speech recognition. Reverbot can’t hear you without both.
• Connect headphones to keep answers private. Without headphones, answers appear on screen instead, unless Speakerphone is on in Settings.
• Have an internet connection, and either free questions left or your own AI key in Settings › AI & API Keys. The count above the big button shows what’s left, for example “3 questions left”.
• Ask in English. For questions in other languages, light Questions beside Translate Vicinity or Translate 2-way.

Tip: To try it on the real screen first, open Help with the ? button and tap Walk me through Questions. The walkthrough is free and uses an example answer.

Ask a question

1. On the main screen, tap Questions under Listening mode.
2. Tap the big button.
3. Wait until “Listening for questions...” appears under the big button.
4. Ask a complete question, such as “What is the boiling point of water?”
5. Stop talking and wait. The answer comes after the pause set in Silence duration before answering.

The status under the Reverbot logo reads “Listening”, then “Processing” while the answer is prepared and “Speaking” while it plays. Reverbot keeps listening for your next question until you tap the big button again.

Listen to the answer

• While an answer plays, tap Tap to skip under the big button to stop it. Reverbot goes back to listening.
• After an answer, tap Repeat that to hear it again.
• With headphones, Reverbot keeps listening while an answer plays, so you can ask your next question at any time; it’s answered next. Without headphones, listening pauses while Reverbot speaks, so it doesn’t hear itself.
• If “Connect headphones to hear it” appears, the answer is on screen instead of spoken. Connect headphones, or tap Speakerphone to hear it out loud. That turns on Speakerphone in Settings, so later answers play out loud too until you turn it off there.

Things to say

Try these in Questions:

• “Who wrote Pride and Prejudice?” — a short, direct answer.
• “Is a whale a mammal?” — yes or no, with the reason.
• “How do I convert Celsius to Fahrenheit?” — how-to questions are answered too.
• “What’s the difference between weather and climate?” — comparisons and explanations.
• “I don’t know what they meant.” — no answer: Questions stays quiet for statements, small talk, opinions and rhetorical questions.

Tip: Name the subject in the question. Questions doesn’t fill in a missing subject, so “How old is it?” gets no answer, but “How old is the Eiffel Tower?” does.

Choose how much detail you hear

Tap the Concise pill under Input & answer to switch to Extended, and tap it again to switch back.

• Concise gives a direct answer in one or two sentences.
• Extended adds useful context, up to four sentences.

The next answer uses the new length, even while Reverbot is listening. The same choice is in Settings, under Answer Style, and it applies to Dictionary too. It changes how much is said, not how fast: voice speed is in Settings › Voice & Speed.

Choose who Reverbot hears

Tap the microphone pill under Input & answer to switch between Room and Solo. In Questions it switches straight away, even while Reverbot is listening.

• Room hears everyone around you, so other people’s questions are answered too.
• Solo isolates your voice with noise cancellation, which helps in loud places. It favors your voice but can’t promise to leave everyone else out.

The same choice is in Settings, under Microphone Input.

Answer questions in other languages

1. On the main screen, tap Translate Vicinity or Translate 2-way.
2. Tap Questions. Both pills light up.
3. Tap the big button.

A question asked in another language is translated into your language first, and answered once the translation has been read. English questions are answered as usual.

The line under the big button says so, for example “Listening — foreign speech is translated to English, then questions are answered...”. To stop the answers, tap Questions again; the translation keeps running. The Answer questions too topic explains the details.

Answer from your documents

1. Touch and hold Questions, then tap Settings.
2. Under Answer from your information, tap Information Source and choose a source.
3. Tap Done, then ask about your material as usual.

Questions then answers with the source’s subject, reference material, answer policy and AI.

With Only my materials, it answers only from passages it can verify, and says so when it finds nothing. With Materials + general AI knowledge, answers can include facts from outside your material, so check them. Choose None — general questions to go back to ordinary answers.

Note: Neither policy searches the web, and the AI still needs an internet connection.

Make an information source

1. Touch and hold Questions, then tap Settings.
2. Under Answer from your information, tap Manage information sources. The Custom Conversations screen opens.
3. Tap the Add button (+), enter a name such as “Handbook”, then tap Add.
4. Under Purpose, set Use this template for to Information Source.
5. Under Information Source, describe the subject, then choose Only my materials or Materials + general AI knowledge.
6. Tap Use with Questions. It changes to Selected for Questions.
7. Under Reference Documents, tap Upload reference files or Write or paste information, and add your material.

Questions answers from the source once its documents are ready. The Custom Conversation topic explains the rest of the editor.

Change Questions settings

Touch and hold Questions, then tap Settings. Or, in Settings, under Listening Mode, tap the sliders button beside Questions. From top to bottom:

• How Questions works opens this Help topic.
• Under Button style, Different button style for Questions gives Questions its own look in the big button, and Style chooses it.
• How long it waits, under Silence duration before answering, sets the pause after you stop talking. If you stop mid-sentence, it waits one second more.
• AI chooses which AI answers Questions, or keeps the default from AI & API Keys. The same AI answers for Answer questions too in other modes.
• Different voice for Questions and Different speed for Questions give answers their own voice and speed.
• Answer from your information chooses an Information Source.
• Restore Questions defaults puts these settings back after you confirm. Your sources and their documents are kept.

What it costs

• Each answer uses one of your free questions while you have them, even an answer you skip partway through. A question that gets no answer costs nothing.
• An AI chosen under AI in Questions settings answers on your own key for it from the start, and that company bills you.
• When your free questions are used, the line under the big button says what’s next: a free trial, a subscription or your own AI key. The subscription topic explains how free questions work.

If it doesn’t work

• If no words appear under the big button, open iPhone Settings, go to Apps › Reverbot and turn on Microphone and Speech Recognition. Then tap the big button twice to restart.
• If your words appear but nothing is answered, ask again as one complete question that names its subject. Statements, small talk and opinions get no answer on purpose.
• If only the first half of a question is answered, drag How long it waits toward Longest in Questions settings.
• If an answer appears but you don’t hear it, check that your headphones are the audio output, or turn on Speakerphone in Settings.
• If an Information Source says it couldn’t find an answer in the available material, ask a more specific question, or check the source’s documents with Edit selected source.

Note: AI answers can be wrong. Check anything important before you rely on it.

Still not working? In Settings, under Feedback, tap Report a Bug and describe what you did, what you expected and what happened.

Before you begin

• When iPhone asks, allow access to the microphone and to speech recognition. Reverbot can’t hear you without both.
• Connect headphones to keep explanations private. Without headphones, they appear on screen instead, unless Speakerphone is on in Settings.
• Have an internet connection, and either free questions left or your own AI key in Settings › AI & API Keys.
• Say the word in English. Dictionary hears English speech.

Tip: To try it on the real screen first, open Help with the ? button and tap Walk me through Dictionary. The walkthrough is free and uses an example definition.

Look up a word

1. On the main screen, tap Dictionary under Listening mode.
2. Tap the big button.
3. Wait until “Listening for words...” appears under the big button.
4. Say a word or phrase, such as “ephemeral” or “the whole nine yards”, then stop talking.

The explanation comes after the pause set in Silence duration before answering. Dictionary then listens for the next word until you tap the big button again. While an explanation plays, tap Tap to skip to stop it; afterward, tap Repeat that to hear it again.

Get the meaning in context

Say the word, then the sentence you heard it in, as one request:

• “What does bank mean in I sat on the bank of the river?” — the edge of a river, not a place for money.
• “Define pitch as used in he made a sales pitch” — the meaning that fits a sales pitch.

Dictionary explains the one meaning that fits the sentence, rather than the most common one. Say the whole request before you pause. You don’t need to say the quotation marks.

Things to say

Try these in Dictionary:

• “Ephemeral” — a word on its own gets its most common meaning.
• “Define ubiquitous” — a request with “define” is always answered, however common the word.
• “How do you spell necessary?” — the spelling.
• “What’s another word for happy?” — a word that means the same.
• “What’s the opposite of generous?” — a word that means the opposite.
• “Where does the word salary come from?” — where the word comes from.
• “How do you pronounce quinoa?” — how to say it.
• “Is irregardless a real word?” — whether it’s a real word, and how it’s used.
• “What is the capital of France?” — no answer: that’s a question for Questions.

Choose how much detail you hear

Tap the Concise pill under Input & answer to switch to Extended, and tap it again to switch back.

• Concise explains a word in one short sentence, with no example.
• Extended adds a short example or usage note, in up to three sentences.

The next explanation uses the new length. The same choice is in Settings, under Answer Style, and it applies to Questions too.

Listen to the explanation

• With headphones, Dictionary keeps listening while an explanation plays, so you can say the next word straight away. Without headphones, listening pauses while Reverbot speaks.
• If “Connect headphones to hear it” appears, the explanation is on screen instead of spoken. Connect headphones, or tap Speakerphone to hear it out loud. That turns on Speakerphone in Settings, so later answers play out loud too until you turn it off there.
• To keep background talk from being looked up, tap the microphone pill under Input & answer to choose Solo, which favors your voice.

Change Dictionary settings

Touch and hold Dictionary, then tap Settings. Or, in Settings, under Listening Mode, tap the sliders button beside Dictionary. From top to bottom:

• How Dictionary works opens this Help topic.
• Under Button style, Different button style for Dictionary gives Dictionary its own look in the big button, and Style chooses it.
• How long it waits, under Silence duration before answering, sets the pause after you stop talking. If you stop mid-sentence, it waits one second more.
• AI chooses which AI explains words, or keeps the default from AI & API Keys.
• Different voice for Dictionary and Different speed for Dictionary give explanations their own voice and speed.
• Restore Dictionary defaults puts these settings back after you confirm.

What it costs

• Each explanation uses one of your free questions while you have them, even one you skip partway through. A request that gets no answer costs nothing.
• An AI chosen under AI in Dictionary settings explains words on your own key for it, and that company bills you.
• When your free questions are used, the line under the big button says what’s next: a free trial, a subscription or your own AI key.

If it doesn’t work

• If no words appear under the big button, open iPhone Settings, go to Apps › Reverbot and turn on Microphone and Speech Recognition. Then tap the big button twice to restart.
• If the word is misheard, say it again as a request, such as “Define serendipity”, or add the sentence you heard it in.
• If nothing is said back, make sure the request is about a word. General questions get no answer in Dictionary; ask them in Questions.
• If Dictionary answers before you finish the sentence you’re quoting, drag How long it waits toward Longest in Dictionary settings.
• If an explanation appears but you don’t hear it, check your headphones, or turn on Speakerphone in Settings.

Note: Explanations are written by AI and can be wrong. Check specialist terms against the source where you found them.

Still not working? In Settings, under Feedback, tap Report a Bug and describe what you did, what you expected and what happened.

Before you begin

• Connect headphones. Translations play in them; without headphones, they appear on screen instead, unless Speakerphone is on in Settings.
• When iPhone asks, allow access to the microphone and to speech recognition.
• Have an internet connection. Each sentence is sent to be translated.
• Have translation available: free questions left, a subscription or your own Gemini or OpenAI key chosen under AI in Vicinity Language Translation settings. A key that only answers questions doesn’t translate.

Tip: To try it on the real screen first, open Help with the ? button and tap Walk me through Vicinity Translation. The walkthrough is free and uses an example translation.

Choose the language you hear

1. Touch and hold Translate Vicinity, then tap Settings. The Vicinity Language Translation screen opens.
2. Under Language, tap Translate to and choose your language.
3. Tap Done.

Translate to lists the 31 languages Reverbot can both hear and speak. The other person’s language isn’t a setting: as “Detects the spoken language automatically” says, Reverbot recognizes it from their speech, which can take a few words.

Note: Translate to is shared with Translate 2-way and Translate too, and with TRANSLATE in the top bar, where it’s called I speak.

Translate what you hear

1. On the main screen, tap Translate Vicinity under Listening mode.
2. Tap the big button. The line under it reads, for example, “Listening — foreign speech is translated to English...”.
3. Hold or place the iPhone where its microphone can hear the speaker clearly.
4. Let the speaker finish a sentence and pause.

The sentence is translated and plays in your language. Reverbot keeps translating, one sentence at a time, until you tap the big button again. The status under the Reverbot logo reads “Processing” while a sentence is translated and “Speaking” while it plays.

What you hear and when

In Translate Vicinity:

• Each sentence is translated after the speaker pauses, not word by word.
• Speech already in your language is skipped on purpose, with nothing said.
• If Reverbot can’t make a sentence out, “Didn’t catch that — say it again” appears for a moment instead.
• With headphones, Reverbot keeps listening while a translation plays, and what’s said meanwhile is translated next. Without headphones, listening pauses while it speaks.
• While a translation plays, tap Tap to skip to stop it. Afterward, tap Repeat that to hear it again.

The microphone is set to Room

Translate Vicinity needs to hear the people around you, so it sets the microphone to Room while it’s chosen. Solo isn’t offered while this mode is chosen: the microphone pill under Input & answer shows Room and doesn’t change, and Microphone Input in Settings can’t be changed either. Your own setting comes back when you choose another mode.

Keep the microphone uncovered and near the person speaking. Room can’t promise to hear someone far away.

Answer questions while translating

To also answer questions asked around you, tap Questions while Translate Vicinity is lit, so both pills light up. A question in another language is translated first, then answered once the translation has been read; English questions are answered directly.

The line under the big button then ends “…then questions are answered...”. Tap Questions again to stop the answers; the translation carries on.

The same switch is Answer questions too in Vicinity Language Translation settings. Each answer is charged like any Questions answer.

Choose who translates

Important: On a free Gemini key, Google’s unpaid terms apply: Google may use the recordings, including other people’s voices, to improve its products, and people may review them, except in the EEA, Switzerland and the UK.

1. Touch and hold Translate Vicinity, then tap Settings.
2. Under AI, tap AI. The AI for Translation screen opens.
3. Choose Reverbot (included), Gemini or ChatGPT.

Reverbot (included) uses your free questions, or your subscription’s translations. Gemini and ChatGPT translate on your own key, and that company bills you; with no key saved for the one you choose, Reverbot translates.

Only these two are offered, because they take the sound itself. The choice applies to Translate 2-way, TRANSLATE and Translate too as well, but phone calls with Both of you always use Reverbot’s translation. It’s also in Settings › AI & API Keys, under AI for each feature, as Translation.

Practice following a language

1. Touch and hold Translate Vicinity, then tap Settings.
2. Under Practice, turn on Rehearse first.
3. Under Partner speaks, choose the language to practice.
4. Tap Done, then tap the big button.

Reverbot plays someone talking nearby in short, everyday sentences, and doesn’t translate them. A banner reads “REHEARSAL — Reverbot is playing the other person, not translating.”

Vicinity rehearsals always use the Standard partner, and each turn costs a question, as an answer does. Turn Rehearse first off to translate for real. The Practice topic explains every rehearsal setting.

Change Vicinity settings

Touch and hold Translate Vicinity, then tap Settings. Or, in Settings, under Listening Mode, tap the sliders button beside Vicinity. The Vicinity Language Translation screen shows, from top to bottom:

• How Vicinity Language Translation works opens this Help topic.
• Under Button style, Different button style for Vicinity translation and Style give translations their own look in the big button.
• Under Language, Translate to is the language you hear.
• How long it waits, under Silence duration before answering, is how long a speaker must pause before the sentence is translated. Shorter is quicker but can cut a sentence into pieces.
• AI chooses who translates.
• Different voice for Vicinity translation and Different speed for Vicinity translation give translations their own voice and speed.
• Under Practice, Rehearse first turns on a rehearsal.
• Answer questions too is the same as lighting Questions beside Translate Vicinity.
• Restore Vicinity Language Translation defaults puts these back after you confirm, including Translate to and the translation AI, which Translate 2-way shares.

If it doesn’t work

• If nothing is translated, check that Translate to isn’t the language being spoken. Speech already in your language is skipped on purpose.
• If “Didn’t catch that — say it again” appears, move the iPhone closer to the speaker and ask people to speak one at a time.
• If sentences arrive in pieces, drag How long it waits toward Longest. If translations come too long after people stop, drag it toward Shortest.
• If translations appear on screen but don’t play, connect headphones or turn on Speakerphone in Settings.
• If translating stops and the line under the big button reads “Subscribe to keep translating”, your free translations are used. Subscribe, or choose your own Gemini or OpenAI key under AI.
• If a subscriber sees “That’s today’s limit on Reverbot’s connection — it comes back tomorrow.”, translation is back the next day; your own key under AI works meanwhile.

Note: Translations can be wrong. Check names, numbers and anything important with the speaker.

Still not working? In Settings, under Feedback, tap Report a Bug and describe what you did, what you expected and what happened.

Before you begin

• Hold or place the iPhone between you, where its microphone hears you both. Translate 2-way sets the microphone to Room for you.
• Headphones are optional. With them, their words, translated for you, play in your ears while your reply comes out of the iPhone’s speaker for them. Without them, everything plays out loud.
• When iPhone asks, allow access to the microphone and to speech recognition.
• Have an internet connection, and translation available: free questions left, a subscription or your own Gemini or OpenAI key chosen under AI in Two-way Language Translation settings.

Tip: To try it on your own first, open Help with the ? button and tap Walk me through Two-way Translation. It can play a Spanish sample for you, and it’s free.

Have a conversation

1. On the main screen, tap Translate 2-way under Listening mode.
2. Tap the big button. The line under it reads “Listening — just talk. Either of you.”
3. Let the other person say a sentence in their language. You hear it in yours.
4. Answer in your language. The phone says your reply in their language, out of its speaker.
5. Carry on talking. Reverbot translates both ways until you tap the big button again.

Reverbot learns their language from what they say first, so it can’t translate your reply until they’ve spoken. While a sentence is being translated, the line reads “Translating...”.

Things to say

Say these during Translate 2-way:

• “Repeat” — plays your last translated reply again for them, at no cost. “Again”, “say that again”, “come again” and “one more time” work too.
• “Sorry, can you say that again?” — the same: polite words around the request are fine.
• “I’d like to repeat my order.” — translated like any sentence. A request to repeat counts only when it’s all you say.
• “Skip” — when you’re saying your replies yourself, moves on to the next sentence.

Choose where your reply plays

With Phone speaks your reply on, your translated reply comes out of the iPhone’s speaker, even with headphones in, so the other person hears it.

• A Speaker pill appears under Input & answer while Translate 2-way is chosen. Tap it to switch to Earpiece when they’d rather hold the phone to their ear, and tap Earpiece to switch back. The change applies to the next sentence.
• The same switch is Out loud on speakerphone in Two-way Language Translation settings.
• It turns back on each time you start Translate 2-way.

Their words, translated for you, play wherever you’re listening: your headphones, or the iPhone’s speaker without them.

Say your reply yourself

1. Touch and hold Translate 2-way, then tap Settings.
2. Under How the conversation runs, turn off Phone speaks your reply.
3. Tap Done, then start Translate 2-way again.

Reverbot now says your reply into your ear in their language, then listens while you say it out loud; the line under the big button reads “Say it out loud, or say ‘repeat’...”. It confirms when you’re close, and after a few tries it moves on anyway. Say “repeat” to hear the phrase again, or “skip” to move on.

Note: Phone speaks your reply and Guide me through it apply the next time you start Translate 2-way. A conversation already running keeps what it started with.

Choose automatic or guided turns

Detect who’s speaking is on at first. Either of you can speak at any time, and Reverbot works out from the sound which way to translate. It doesn’t identify anyone, and in a noisy room it can get it wrong.

Turn it off in Two-way Language Translation settings to take turns instead: they speak, you hear the translation, then you reply. Speech out of turn is skipped.

The line under the big button reads “Listening for them...” during their turn and “Your turn — say what you want translated...” during yours. With Guide me through it on, Reverbot also says it, for example “After the beep, say what you’d like to translate.”

This switch takes effect straight away, even mid-conversation. On a phone call, Reverbot always detects who’s speaking.

Answer questions while translating

Tap Questions while Translate 2-way is lit, so both pills light up. Reverbot then also answers questions asked in the conversation: theirs after their translation has been read to you, and yours directly. The line under the big button reads “Listening — just talk. Their questions are answered after they’re translated.”

Answers wait for a gap and never talk over a translation.

Tap Questions again to stop them; the conversation carries on. The same switch is Answer questions too in Two-way Language Translation settings, and each answer is charged like any Questions answer.

Practice a conversation

Each Standard turn costs a question. A live partner is billed by the minute, and your voice streams to that company while it runs.

1. Touch and hold Translate 2-way, then tap Settings.
2. Under Practice, turn on Rehearse first.
3. Under Partner speaks, choose the language to practice.
4. Under Partner’s voice, choose Standard, Live — Reverbot or Live — OpenAI (your key).
5. Tap Done, then tap the big button.

Reverbot plays a native speaker who greets you and asks simple everyday questions, and doesn’t translate. Standard takes turns; a live partner talks back and forth and lets you cut in.

Turn Rehearse first off to translate for real. The Practice topic explains every setting.

Translate a phone call

Translate 2-way also works on a phone call you place from Reverbot: choose Both of you, then Two-way translation. The Call a phone number topic explains placing the call.

Important: The person you call hears an automated announcement that AI is being used, and must press 1 before Reverbot listens to them. It plays on every Both of you call, and you confirm twice before each one.

On a call, their words reach you in your language and your reply goes down the line in theirs. Reverbot always detects who’s speaking, and always uses its own translation, never your key.

Change Two-way settings

Touch and hold Translate 2-way, then tap Settings. Or, in Settings, under Listening Mode, tap the sliders button beside Two-way. The Two-way Language Translation screen shows, from top to bottom:

• How Two-way Language Translation works opens this Help topic.
• Under Button style, Different button style for Two-way translation and Style give Two-way its own look in the big button.
• Under Language, Translate to is your language, shared with Translate Vicinity.
• Under How the conversation runs: Phone speaks your reply, Out loud on speakerphone, Guide me through it and Detect who’s speaking. Guide me through it says what to do at each step.
• How long it waits, under Silence duration before answering, is the pause that ends a sentence, for either of you. TRANSLATE in the top bar uses it too.
• AI chooses who translates, shared with Translate Vicinity.
• Different voice for Two-way translation and Different speed for Two-way translation give translations their own voice and speed.
• Practice holds Rehearse first, and Answer questions too matches the Questions pill.
• Restore Two-way Language Translation defaults turns the four switches back on and resets the rest, including Translate to, after you confirm.

If it doesn’t work

• If their words aren’t translated for you, check that Translate to is your language, not theirs.
• If your reply isn’t translated, let them speak first: Reverbot learns their language from them.
• If your own words are played back to you, or theirs go out of the speaker, turn off Detect who’s speaking to take turns instead, and keep the iPhone a little closer to you.
• If they can’t hear your reply, check the iPhone’s volume and the pill under Input & answer. If it reads Earpiece, tap it to switch back to Speaker.
• If sentences are cut into pieces, drag How long it waits toward Longest.
• If you say your replies yourself and Reverbot never confirms them, your iPhone may not recognize their language; it then moves on after each phrase.
• If translating stops and the line under the big button reads “Subscribe to keep translating”, your free translations are used. Subscribe, or choose your own key under AI.

Still not working? In Settings, under Feedback, tap Report a Bug and describe what you did, what you expected and what happened.

Before you begin

• Everyone needs Reverbot open on the Translate nearby screen: tap TRANSLATE in the top bar. Reverbot can only find phones that are running it.
• Turn on Wi-Fi and Bluetooth, and allow Local Network when iPhone asks. The phones connect to each other directly.
• Each phone needs an internet connection for translation, and translation available: free questions, a subscription or that person’s own Gemini or OpenAI key.
• If you haven’t agreed before, Before we start appears: turn on I’m 18 or older, then tap Agree and continue. What you say is sent to an AI to be translated.
• Each of you can wear your own headphones. Keep Reverbot open during the conversation: the screen stays on, and locking the phone or leaving the app can break the connection.

Your name and photo

1. Tap TRANSLATE in the top bar.
2. Your name card at the top shows what others see. Tap Edit. The Profile screen opens.
3. Under Name, enter the name people nearby see, up to 40 characters. Left blank, it uses your Reverbot username, or else your account name or the iPhone’s name.
4. Optional: Under Photo, tap Add a photo and choose one.
5. If you added a photo, turn on Show my photo. If not, choose a picture under Avatar.
6. Check How you appear at the top, then tap Done.

How you appear is exactly what the people you connect with see. Remove photo deletes the photo from this iPhone, but not a copy someone already received.

Find each other and connect

1. Both of you tap TRANSLATE in the top bar.
2. Under Your language, each of you sets I speak to your own language.
3. Wait for the other person to appear under Nearby, with the language they speak.
4. One of you taps Connect beside the other’s name. It reads “Connecting…” while it works; don’t tap again.
5. When the screen changes to Talking and shows “Live”, start talking.

I speak is the same setting as Translate to in Translate Vicinity and Translate 2-way. Changing it, or Group (more than two people), restarts the search.

Talk and follow the conversation

• Speak into your own phone, a sentence at a time, and pause between sentences. Your phone listens with Solo, so it favors your voice.
• Your phone translates what you say into each other person’s language and sends it to them. Theirs does the same for you.
• Their lines show the translation first, with their own words under it, and are read out to you. Your lines show what you said, with what they heard under it.
• While a line plays, a banner shows whose it is and how many are waiting.
• Tap Repeat at the bottom to hear the last line again.
• To finish, tap End at the top of the screen.

The bottom of the screen reads “Just talk — they hear you in their language.”, and “Translating…” while your sentence is on its way.

Talk in a group

1. Before anyone connects, each person who should be able to bring others in turns on Group (more than two people) under Your language.
2. Under Nearby, tap Add beside each person you want in the conversation.
3. Once you’re talking, anyone else in range appears beside Add, under the people at the top. Tap a name to bring them in.

With Group (more than two people) off, a phone connects to one person: it can join a group through someone else, but nobody can join through it. Each sentence you say is translated once for every other language in the group.

Answer questions in the conversation

On the Translate nearby screen, turn on Questions under Your language before you connect, or, once you’re talking, tap the Questions pill under the people; it shows On or Off. Reverbot then answers questions asked in the conversation, in your ear only: theirs once they’ve been read to you in your language, and yours once they’ve gone out. Each answer counts as a question, as any answer does.

Share your contact card

1. On the Translate nearby screen, tap Edit on your name card.
2. Under Contact card, tap Choose from Contacts and pick your card. My Card is at the top of the list.
3. Tap What they get and turn on anything you want to share beyond your name and phone numbers.
4. Tap Done.
5. In the conversation, tap Card at the bottom.
6. Check who it goes to, then tap Send my card.

Your card goes only to the people in the conversation, and only when you tap Card. They can tap Add to Contacts or Share. On the Profile screen, Show as QR code lets anyone save your card with their camera, without Reverbot, and Share card… sends it as a file.

Finish your profile

The Profile screen, opened with Edit on your name card, also holds:

• Username: how friends find you on Reverbot. You can change it once every 60 days, free, if the new one is available.
• Work: an optional Company and Job title, up to 60 characters each, shown under your name in nearby meetings and used by a meeting’s report to tell people apart.
• Avatar: a picture, color and initials font, used when Show my photo is off or there’s no photo.
• Friends: Friends see my name and Friends see my photo decide what your Reverbot friends see besides your username.

When you’re signed in, the same screen opens from Settings: under Profile, tap your name.

Messages and recording requests

During a nearby conversation:

• Messages appears under the people once someone sends you a message, for example from an AI Meeting’s room. Tap it to open the chat; the number shows how many are unread.
• If someone starts an AI Meeting with you in the conversation, a request appears on your phone, for example “Sam wants to record this meeting”. Tap I agree to be recorded and transcribed, or No. You can withdraw at any time with Withdraw, and the recording is deleted.
• If someone asks for your attention from their meeting screen, their name and photo appear over the conversation.

Change how it sounds

• The voice that reads other people’s lines: in Settings › Voice & Speed › Voice for each mode, under Translation, turn on Nearby (TRANSLATE) and choose a voice.
• Its speed: in Settings › Voice & Speed › Speed for each mode, turn on Nearby (TRANSLATE) the same way.
• The pause that ends your sentence: How long it waits in Two-way Language Translation settings, which TRANSLATE shares.
• I speak, Group (more than two people) and Questions are remembered for next time.

What leaves your iPhone

• The phones talk to each other directly, over an encrypted connection: your name, language, photo or avatar, work details, the translated lines and your contact card when you send it.
• What you say is sent over the internet to be translated: to Reverbot’s translation service, or to Google or OpenAI when you chose your own key for translation.
• Nothing of the conversation is kept: its lines and messages are cleared when it ends.
• Each sentence you say is translated once for each other language present. Before you subscribe, each translation uses one of your free questions; with a subscription, it comes from a daily allowance.

Note: An AI Meeting with people in the room keeps its own transcript and recording, under its own rules.

If it doesn’t work

• If nobody appears under Nearby, check that the other phone has Reverbot open on the Translate nearby screen. After a while a hint appears with Open Settings: check that Local Network is on for Reverbot.
• If an orange message appears instead of the search, turn on Wi-Fi and Bluetooth.
• If “Reconnecting…” appears, bring both phones back to Reverbot and keep them close. If the conversation ends, the reason shows on the Translate nearby screen; connect again.
• If you’re connected but nothing is translated, read the message at the bottom, such as “Couldn’t reach translation — check your connection.” or “You’ve used your free questions — subscribe to keep translating.”, which comes with Subscribe. A subscriber can see “That’s all the translation for today — it comes back tomorrow.”
• If it reads “You all speak English — nothing to translate.”, everyone chose the same language under I speak.
• If it reads “Reverbot can’t hear you — allow Microphone and Speech Recognition in Settings.”, turn both on in iPhone Settings, under Apps › Reverbot.

Still not working? In Settings, under Feedback, tap Report a Bug and describe what you did, what you expected and what happened.

Before you begin

• Both people need a Reverbot account, an internet connection and microphone access.
• Have the other person’s exact username. A phone number is not used for a Reverbot Call.
• Check where audio should play: headphones, earpiece or speaker.
• There is no Reverbot per-minute charge for this call; your mobile data plan may apply.

Make your first call

1. Tap the phone button on the main screen. The Call screen opens on Reverbot Call.
2. Enter their name in Search by username and select the matching account. Check the username before calling.
3. Tap the green phone button. You do not have to become friends first.
4. Wait for them to accept. Speak normally once the call connects.

Find each other again

1. Open Settings and find Profile. Your username appears there with an @ sign.
2. Share that exact username with the person you want to call.
3. On Reverbot Call, search for their username and check the match.

Add them as a friend if you want to find them again without searching. Review a friend request before accepting it.

Answer, change audio and hang up

An incoming call offers Accept and Decline, including on the lock screen. During a call, use Mute to stop sending your microphone and the audio-route control to choose your headphones, earpiece or speaker. Tap End to disconnect. If a second call arrives while you are already in one, the second caller is told you are busy.

Choose what you share

Review My name and My photo on the call screen. These control profile information, not whether your voice is transmitted. Non-friends initially see your username; available profile-sharing choices depend on your profile and the relationship. Friends also see the voice display you chose for your profile wherever you don’t share a photo with them; others don’t. Send my contact card is a separate, deliberate action. Only share details you intend the other person to keep.

Manage recent callers and blocks

Use Recent to find a previous call. Touch and hold a recent entry to delete it from this iPhone. Swipe a person’s row to find Block; review the Blocked section to unblock them. Blocking prevents that account from calling, requesting friendship or finding you, without sending them a block notification.

Privacy and limits

Ordinary Reverbot Call audio is encrypted between the phones; a relay can carry the encrypted traffic when needed. This call is not a recording. Recent-call history is local to this iPhone and excluded from backups. AI-enabled or recorded meetings have separate notices and consent steps; do not assume an ordinary voice call has enabled those features.

If it doesn’t work

• If a person does not appear, check their exact username and both phones’ internet connections.
• If the other person cannot hear you, check Mute and Reverbot’s Microphone permission in iPhone Settings.
• If you cannot hear them, use the call’s route control. Being paired in Bluetooth does not mean this call is playing through your headphones.
• If a call stays connecting, end it and try again once both connections are stable.
• If you are already on a call, a second caller is told you are busy.

Still not working? In Settings, under Feedback, tap Report a Bug and describe which part of connecting or audio failed.

Before you begin

Open the round Chat button on the main screen and let protected storage and encryption finish setting up. Unlock the iPhone if Chat says it cannot open storage. If prompted to use Chat on this iPhone, read that account/device prompt before continuing. You need an internet connection to deliver new messages.

Start a conversation

1. In Chat, tap the pencil/New chat button.
2. Select a friend, or enter an exact username in Search usernames and submit Search.
3. Open the person’s chat, enter text in Message and tap Send.
4. To make a group, choose New group, name it, choose at least two people and tap Create. For example, create “Saturday plans” and send the meeting time to the group.

Send attachments and voice notes

1. Open a conversation in Chat.
2. Tap the Add button (+), then choose Photo or File. Text in Message, up to 2,000 characters, becomes the caption.
3. For a voice note, hold the microphone while speaking and release it to send. Slide left before releasing to cancel.
4. Tap a downloaded attachment to open it. Files and attachments from non-friends require a download tap first.

A voice note can be up to 5 minutes. Stop Reverbot’s listening session or call before recording or playing one.

Important: Each attachment is encrypted on your iPhone before upload and can be up to 25 MB. The server keeps its sealed copy for 30 days; download anything you want to keep. Touch and hold an attachment for sharing and deletion options.

Verify the other person

1. Tap the conversation header to open Chat info.
2. Tap Safety number.
3. Compare the number with the other person in person or over a trusted call, or scan their code.
4. Tap Mark as verified only after the numbers match.

A new phone or account sign-in can change the number. Review the security-change notice before accepting new keys or sending sensitive information.

Set message retention

Open Chat info › Disappearing messages and choose the timer. It applies to the conversation’s participants. Touch and hold a message and choose Delete… for the available deletion actions. Delete for me removes your local copy; Delete for everyone is available for your eligible recent messages, within the app’s 24-hour window. People can still copy or screenshot content before it disappears.

Start a call or video meeting

Use the phone or camera button in the chat. The person starting a Chat call or video meeting must first complete the Cloudflare setup in Chat settings; invitees join the resulting session. A video call or video meeting takes up to 8 people, you included; in a bigger group, start a voice call instead. Screen sharing broadcasts the whole screen, including notifications, so prepare a Focus or hide sensitive content before starting it. Use the call’s end/leave control when finished.

Manage contacts and notifications

Open Contacts to review friend requests or use the optional contact finder. Allow contact access only if you want that lookup; being discoverable by your phone number is a separate opt-in. In Chat info, Mute silences message notifications but does not stop calls ringing. Pin keeps a conversation easier to find. Clear messages deletes this iPhone’s messages while leaving the chat and other people’s copies.

Storage, AI and recordings

Chat messages stay in protected local storage and are excluded from iCloud and computer backups; reinstalling or moving phones is not a chat-backup strategy. The server carries encrypted envelopes and removes delivered envelopes, or undelivered ones after the retention period shown in Chat settings. A photo, voice note or file stays on the server only as its sealed copy, for 30 days after it is sent, or until its sender deletes it for everyone. An AI feature or recorded meeting has its own disclosure and agreement: enabled AI can send relevant content to its AI provider. Do not treat that provider processing as ordinary private messaging.

If it doesn’t work

• If a message shows a clock, it is still sending. Check connectivity before retrying a failed message.
• If a safety-number warning appears, verify the change with the other person. Do not accept it solely to dismiss the warning.
• If an attachment fails, check its size and the message about storage. Files over 25 MB cannot be sent; space also frees as older sealed attachments expire.
• If a voice note cannot record or play, stop the active Reverbot listening session or call first.
• If a call cannot start, check the required Cloudflare setup in Chat settings.
• If protected storage is unavailable, unlock this iPhone and reopen Chat. Do not reinstall to troubleshoot missing local history.

Set up phone calls

Important: Calls use prepaid funds. Review the displayed price before purchasing or dialing. Caller-ID verification can use up to one minute of calling credit.

1. Tap the phone button on the main screen.
2. Tap Phone Call, then sign in if prompted. A guest account cannot hold a calling balance.
3. Turn on Phone calls and read the charges and terms.
4. Tap Add funds and review Apple’s purchase sheet.
5. Complete verification of your own number: answer the verification call and enter its code.

Your verified number is used as caller ID. Do not verify someone else’s number.

Choose who Reverbot can hear

On Call a number, choose Just you or Both of you before selecting the feature. Just you sends only your side into Reverbot’s listening features; the other person’s voice is not supplied to that AI path. Features that require their side are unavailable there. Both of you permits those features only after the recipient hears the automated notice and presses 1.

Understand the announcement

Important: The person you call WILL hear an automated message about AI listening and must press 1 before Reverbot hears their side. This includes someone who does not use Reverbot.

The announcement applies to every Both of you call and every feature offered there, including Questions, Flirt, translation, coaching, Information Source and AI Meeting. Check the recipient carefully before proceeding; it may be inappropriate to surprise someone with this notice.

Reverbot requires this consent flow throughout the US. A one-party-consent state, a location setting or GPS confirmation does not turn it off. This is the service’s mandatory policy for its supported call flows; the app offers no bypass.

Dial and confirm the recipient

1. Open Call a number and enter a supported number, or tap Choose from Contacts.
2. Check the recipient and caller ID.
3. Under What Reverbot can hear, choose Just you or Both of you.
4. Choose the feature under What Reverbot does, and the recipient’s language when offered.
5. Read the displayed price and start the call.
6. For Both of you, read the first warning and tap I understand they WILL hear the message.
7. On the final warning, confirm I checked the recipient and want them to receive this permission announcement., then tap Yes, call and play the announcement.

Both confirmations are required each time. Wait for the recipient to press 1 before expecting assistance that hears them.

Use the feature during the call

For a private dictionary check, choose Just you with Dictionary and say the word yourself. For a two-sided conversation, choose Both of you and the appropriate feature before dialing. To answer from your own reference material, use an Information Source: with Just you, Questions answers what you ask from the source selected for Questions, or from the active source when you start the call with its Custom Conversation pill selected; with Both of you, choose Information Source: once they press 1, what they ask is answered from it, in your ear. Keep Just you calls on the earpiece or headphones: Reverbot pauses listening on loudspeaker to avoid hearing the far-end voice through the phone microphone. Use the call controls to mute, change route, send keypad digits or end the call.

Consent, recording and stopping

Both of you calls play a recurring reminder tone. Recipient permission to AI listening is required before that listening starts; AI Meeting additionally records under its recording flow. Other listening features do not turn the call into a saved audio recording. A recipient can object with “stop listening”; follow the resulting on-screen restriction. Hang up using the call’s end control whenever you want to stop.

Balances and failed calls

Review the current rate on the calling screen rather than assuming every destination costs the same. Calling starts billing when the recipient answers, including time spent on the permission flow. The app warns when credit is running low and ends the call when the balance cannot continue. If dialing is unavailable, check sign-in, Phone calls, verified caller ID, funds and the exact error; server availability can also limit the service.

Emergency and service numbers

Use the iPhone’s ordinary Phone app for emergencies. Reverbot does not place emergency or short service-code calls; supported handoff opens the system dialer. A funded Reverbot account is not a replacement for emergency calling.

If it doesn’t work

• If calling is unavailable, check sign-in, Phone calls, verified caller ID and Your balance.
• If a number is rejected, use a supported US or Canadian number. Reverbot does not place emergency or short service-code calls; use the iPhone’s Phone app for those.
• If assistance is waiting on a Both of you call, the recipient must finish the announcement and press 1. Refusal cannot be bypassed.
• If Just you pauses listening, choose the earpiece or headphones. Loudspeaker playback could let the microphone hear the other person.
• If someone cannot hear you, check Mute and the call’s audio route. End a stalled call and retry after your connection recovers.

Still not working? In Settings, under Feedback, tap Report a Bug. Do not include private call content.

Before you begin

• Souls are for adults 18 and over. Read the introduction before continuing.
• Have an internet connection and an available AI allowance or your own configured key. Advanced generation and paid media can require your own provider account.
• A soul based on a person is an AI representation, not that person. Use real faces and voices only with the required permission.
• Reverbot shows an AI notice at the start of the day’s first conversation and after every three hours of talking, whatever character the soul plays.

Note: Souls are companions — not therapy, counseling or medical care. If you’re in crisis, call or text 988 in the US and Canada, or find a helpline in your country at findahelpline.com. In the US and Canada, call or text 988 for crisis support.

Make a soul

1. Tap SOULS in the top bar, then Make a soul.
2. Choose a starter, such as Trash talker, or use the available custom or AI creation choice.
3. Review the name, look and voice. Use Another look, Edit look or Another voice where offered.
4. Tap Create, or Create [name] when a name is entered.
5. Open Personality & story to edit the character, background, speaking style and custom wake phrases.

On the first profile prompt, Set up my profile helps souls ask to learn about you; Skip for now leaves that for later. A soul receives the profile only after permission.

Call them or use Chat

1. Open the soul’s page and turn on Answers to their name.
2. Read its wake phrase, then close the page and return to the main screen.
3. Tap the big button and say the phrase, such as “Hey Sally, you there?”
4. Talk after the soul joins. Tap the big button to stop listening when finished.

For text, open Chat on the soul’s page. Show calls in Chat includes voice conversations there; Conversations lists saved sessions. Voice conversation starts from the main screen, not the settings page.

A soul never starts a new conversation on its own; you call it.

In Souls settings, preferences apply to all souls:

• Pauses allows occasional explained pauses. Ask “What’s up?” for a check or say “Go on” to continue. In Chat, an explained pause can split a reply when Natural texting pace is on.
• Keep talking in a lull lets a soul occasionally carry on while an existing call is quiet. Extra AI requests can cost money; Reverbot’s AI limits them to two per call, including check-ins and sign-off.
• Thinking sounds adds a quiet sound during a longer wait. Expressive sentence delivery uses the smart response allowance to choose tone without changing the selected voice.

Teach facts and background

1. Open the soul’s page. Under Teach [name], enter one Fact or background detail to add and tap the arrow to add it.
2. For source material, open Documents & facts and add a document or use Paste text.
3. Choose AI that studies them, explain what to learn if needed and use Study them. Review the cost before analysis.
4. Use Life story › Add a moment from their life for a dated event.

A conversation never rewrites an explicit profile entry. Souls can remember conversations, what they learn about you and fictional details they introduce, but the profile material you deliberately enter takes precedence. Apply any suggested personality changes separately.

Review conversational memories in What they remember, and transcripts in Conversations. Forget everything about me removes the relationship memories listed in its warning while keeping profile facts and transcripts; Undo all growth has its own reset confirmation.

Record about them

1. Stop any listening session or call.
2. On the soul’s page, tap the big red Record about [name] button, for example Record about Sally.
3. Speak about their background, people or experiences, then tap Stop recording.
4. Correct Review transcript.
5. Tap Keep just the words to save text without AI sorting, or Add to [name] to review the cost and let the selected AI sort facts and life events.
6. Use Record more for another take, or Discard to leave without keeping it.

This recorder keeps the words, not original microphone audio. To preserve original audio, use Life Journal’s Record a memory. Imported journal recordings can later offer Play original recording on the soul’s Recordings page.

Choose their voice and look

1. Open the soul’s AI & voice page to choose who writes their lines, their voice engine and speed. Add any required provider key before choosing a paid voice.
2. Use Record a voice for this Soul only with the speaker’s permission and an eligible ElevenLabs account.
3. Open their look editor and choose a character, permitted picture or voice display. Tap Done to save; Cancel keeps the previous choices.

The center shows the soul’s chosen voice display, not its profile photo. Talking faces and lips use the configured display; paid realistic face animation needs its own account and consent. A preview in settings does not begin a conversation with the soul.

Set growth and relationship pace

Open Growth & journal on the soul’s page:

• Difficulty, under Relationship difficulty, controls how gradually familiarity and trust develop. Personality default uses a natural pace for most souls and a slower pace for guarded characters. Time and repeated conversation both matter; asking to be closer does not bypass the limits.
• Grows from your conversations allows reflection after conversations.
• How fast controls personality adjustments, not voice speed or the relationship’s time requirements.
• How far from what you set limits drift from your chosen personality.

Keeping to yourself slows closeness. Trust and closeness can go backwards after difficult interactions. The soul’s Journal records its reflections; it is separate from your private Life Journal.

Choose their life and permissions

Under Their life, choose In your phone or describe A life of their own. Realism adds AI-written experiences and moods between conversations; review its own-key requirements and cost before turning it on. None of those fictional events actually happened.

In Permissions, allow only the capabilities you want. Full access allows the listed capabilities, including future additions, but does not bypass iPhone permissions or action confirmations. Make [name] your assistant assigns the one assistant role and organizer access.

Important: Only a soul that lives in your phone and knows it is an AI can have phone access or be your assistant. A soul with a life of its own, one configured to believe it is alive, or a Life Journal narrator cannot use those capabilities. Messages remain drafts you review and send.

Their world: see their day

1. Open a fictional soul’s Their world page and review Your BytePlus account.
2. Set up your own key, then review the face and video permissions.
3. During a conversation about its fictional day, ask “Show me a video of what happened.” Tap Show me when offered.
4. Tap Write the scene and review the scene. Nothing has been filmed or generated yet.
5. Review the separate video price, quality and length, then confirm Make the video.

The AI-made result is kept in Their world and can be watched again. These are fictional scenes, not recordings of real events. BytePlus limits which faces its video service accepts; read the face message before paying for generation.

Create photos and videos

Important: Media generation is optional for each soul and uses the configured provider’s account. Image and video requests are separate charges.

1. Open AI & voice › Photo & video creation.
2. Allow images and videos separately for this soul.
3. Choose image and video providers, or follow the account defaults from Settings › Image & Video AI.
4. Allow on-screen display after confirmation only if wanted.
5. Describe a non-explicit image or video, then review the request and confirm creation.

Settings › Media library holds creations, saved jobs, favorites, sharing and deletion. Nano Banana is an image model; optional assistance can prepare image references for supported video services or edit a journal image as a copy. It does not generate audio or replace the original source.

Introduce souls and share deliberately

You can call up to four souls into a get-together from the main screen. Invite another by name or ask those present to discuss a subject for a round. Their own AI and voices are used. On a soul’s page, Souls they know lets you introduce them and review their conversations with each other.

Important: Sending or saving a .soul file moves the soul off this phone; it is not an ordinary backup copy. Whoever opens the encrypted file first becomes its owner, and other copies will not open.

To share, turn on Export, import & send souls in Souls settings, then use Send or save as a .soul file or Send to a friend on the soul’s page. Read the media-storage choices and password option before confirming.

Get help using Reverbot

1. Tap SOULS, then Make a soul.
2. Choose Reverbot App Helper at the start of the starter list.
3. Add your own supported AI key if needed. This helper does not spend Reverbot’s free questions and starts with the iPhone’s on-device voice.
4. Ask for help, such as “Where do I change the voice for Questions?”

The helper uses Reverbot’s current written Help, opens supported settings pages and can point to controls the app identifies. It does not inspect screenshots or take over the phone. You still choose settings, grant permissions, enter keys and confirm calls, messages, purchases or deletion yourself.

You can give another eligible soul this ability in Permissions, using Guide you around Reverbot. A soul with a life outside the phone cannot receive this permission.

If it doesn’t work

• If a wake phrase does not answer, turn on Answers to their name, start listening with the big button and check the phrase. Duplicate phrases need renaming.
• If replies fail, check AI & voice, its key, connection and allowance.
• If the voice differs, check the soul’s voice engine and any missing-key message.
• If a permission is unavailable, review Their life and whether the soul knows it is an AI; fictional lives outside the phone cannot control it.
• If a paid scene fails, read Couldn’t be made — why and check the provider’s face and region requirements before retrying.
• If a remembered fact is wrong, correct its profile source instead of repeating it in Chat.

Still not working? In Settings, under Feedback, tap Report a Bug.

Turn on Life Journal

1. Open Settings › Life Journal.
2. Turn on Enable Life Journal. It is off by default.
3. Sign in if asked, so the journal belongs to your account.
4. Under Add a memory, choose Record a memory, Write a memory or an import.

Turning on Life Journal does not start background recording or grant every permission. Each personal source, AI request, backup and family package has its own control.

Record and save your own voice

1. Stop listening or any call, then open Settings › Life Journal.
2. Tap Record a memory and enter a Memory title, such as “My first day at school”.
3. Tap Record memory and allow microphone access if asked. Record yourself or someone who agreed to recording; the limit is 30 minutes.
4. Tap Stop recording, then Preview recording.
5. Optional: Add Notes or transcript.
6. Tap Save.

The recording appears in Your timeline. Stopping alone does not save it, and saving needs a title. Keep the recorder open until saved; Cancel discards the temporary take.

Play, transcribe and correct a recording

Open the saved timeline entry and tap Play saved recording. To make text, choose Transcribe on this phone and allow Apple Speech permission when requested. This transcribes the existing file, without reopening the microphone or falling back to a server. On-device recognition must support the recording’s language; partial or incorrect text needs review. Use Edit entry to correct it before marking Reviewed and approved for drafts.

Write or import selected material

1. In Settings › Life Journal, tap Write a memory to enter a title, date, chapter and text, then tap Save.
2. To import media, tap Choose photos or videos and select the items in the system picker.
3. For documents, tap Import documents or files. Google Drive, Dropbox and other services appear through Files when installed and enabled there.
4. Optional: Use Add from Journaling Suggestions where available to choose a moment offered in Apple’s window.

Only the material you choose is imported. A folder can also be tracked under Choose source permissions.

Organize new imports with AI starts off. When on, each later imported photo, video or document is sent once to the chosen AI for suggested titles, dates, chapters, people, places and facts. Review AI suggestions to review before using those suggestions. A video sends three still frames, without sound. Recordings and daily source entries are not automatically sent.

Important: Organizing an import uses the displayed own-key request or free-question allowance. Turning on automatic organization authorizes those subsequent requests.

Choose personal sources

1. Open Settings › Life Journal › Choose source permissions.
2. Turn on only the sources you want.
3. Read the explanation and allow the matching iPhone permission when asked.
4. Review Added from sources and the new timeline entries.

• Calendar, Reminders and Contacts add the selected information for review. Calendar can work further back gradually.
• Photo details records dates, locations and on-device image descriptions, not unrestricted access to every app. Place-name lookup sends the location to Apple Maps.
• Current location reads your location when you turn it on or refresh. It replaces the previous current-location entry unless you edited that entry yourself.
• Places you visit needs Always location permission so iPhone can notice visits while Reverbot is closed.
• Motion & fitness is read-only; those daily entries never become soul sources.
• A folder you choose imports supported files in that folder and its subfolders, up to 20 new files per pass. Unreadable files are set aside until changed.

Sources normally add one entry per source and day while Reverbot is open; permitted background refresh may also run while the phone is unlocked. A deleted day is not added again. Source entries are not sent automatically to AI or included in drafts before approval.

Use a private listener interview

Important: The listener’s AI receives what you say during the interview. Without headphones, the listener speaks aloud.

1. In Settings › Life Journal, choose Simple guide or a soul under Your listener.
2. Tap Start a private interview, or say the Life Journal wake phrase during an active listening session.
3. Tell one memory at a time. Try “Ask me about my childhood home, one question at a time.”
4. Pause for the response; Wait before your listener answers sets the pause. Additional words while it is thinking join the same turn.
5. Tap Stop interview, then Save interview to put it in the timeline, or Discard.

If you say the phrase while talking with a soul, that soul switches into listener mode. Its stored memories and older conversations are excluded; only the last few lines of the conversation you switched from accompany the handoff.

Also let this Soul remember saved interviews starts off. Turn it on before the interview and leave it on when saving only if the listening soul should retain the saved interview. Simple guide uses Listener display & voice; a soul uses its own voice.

Set the listener wake phrases

1. In Settings › Life Journal, turn on Use the Life Journal wake phrase under Start an interview by voice.
2. Read Main wake phrase. Its default is “Hey Life Journal Listener”.
3. Optional: Enter a phrase with at least two words. Until it is valid, the default remains in use and the field explains why.
4. Optional: Turn on Also listen for “Hey Journal”. It starts off. If the main phrase is exactly “Hey Journal”, ignoring capitalization, this duplicate switch is unavailable.
5. Return to the main screen and start listening with the big button before saying the phrase.

You can also find Life Journal Listener under Wake Phrases in Settings after Life Journal is enabled. Calling a soul by name starts an ordinary soul conversation; say the journal phrase during it to switch into the private listener interview.

Analyze sources and create a draft

Important: Source analysis and draft generation send the selected material to the AI shown for that request. Review the confirmation and cost first.

1. Open a timeline entry and tap Analyze this source.
2. Review the proposed facts and correct errors in Edit entry.
3. Turn on Reviewed and approved for drafts only when you accept the source.
4. In the timeline, select Include for each approved source you want.
5. Under Life story drafts, choose the AI and tap Generate from selected approved sources.
6. Check the draft against its linked sources, edit it and tap Save edits.

Unreviewed AI guesses should not become biography. The linked sources let you check where an assertion came from.

Back up and recover

Your journal is encrypted on this iPhone with its own key before backup. Reverbot’s server stores locked copies it cannot read; backup storage is included up to the displayed account limit.

1. Turn on Back up automatically for changed pieces to upload after about a minute and when you open Reverbot, or tap Back up encrypted journal now.
2. After your first backup, tap Save recovery kit when prompted.
3. Keep that kit somewhere private and recoverable. Reverbot cannot recover its decryption key for you.

Important: The recovery kit contains the key and is not password-sealed. Do not send it as an ordinary journal attachment.

If another phone made the server backup, choose Restore that backup or Replace it with this iPhone’s journal only after reading the replacement warning. Replacement deletes the other snapshot. Restore recovery kit and Restore encrypted cloud snapshot also explain what they replace before continuing.

Leave memories for family

1. Open Settings › Life Journal › Leave memories for family and tap Leave memories for someone.
2. Choose them from your friends, or type their Reverbot username. If they haven’t turned on receiving yet, the package waits on your iPhone and is sealed to them as soon as they do.
3. Optionally give their email and turn on Tell them now that I chose them — both need a verified email address on your own sign-in; otherwise they hear nothing until something is released to them.
4. Choose what they receive: the memories you pick, or Everything I approve, now and later — and, if you want, your Journal Soul, your recorded face and voice, and life story drafts made only from those memories.
5. Type the name they’ll see and an optional message, which is encrypted. Choose a waiting period of 7 to 365 days, and leave Keep this package up to date on if it should follow your journal.
6. Tap Set aside, then check the encryption fingerprint with them.

Nothing is released while you’re alive, and inactivity never releases anything. Someone you chose must report your death with an official record; you’re told at once — by email if your sign-in has a verified address, and in Life Journal — so you can choose I’m alive — stop this; two independent reviewers must confirm it; then your waiting period passes. Only then is it released, and they’re told by email. Revoke a package at any time; copies already released can’t be taken back.

Record your face and test your Journal Soul

1. Turn on Life Journal, then open My Journal Soul, face & voice.
2. Enter Your journal Soul’s name, then tap Save name.
3. Tap Record my face & permission and follow the recorded consent flow. Sharing and use after death are separate permissions.
4. Select the approved memories to include, then tap Create Journal Soul.
5. Open AI, voice & recorded voice setup to choose the AI and voice. A recorded face has free local animation; Set up face animation offers optional paid services on your own account.
6. Tap Test with a text conversation to ask about the saved material. For voice, turn on Answers to their name, return to the main screen, tap the big button and say the soul’s phrase.
7. Correct biography in journal entries or profile sources. Test conversations do not rewrite those sources.
8. Optional: Turn on Keep my Journal Soul up to date, or tap Update now. Approved entries are included and removed or unapproved sources are taken out. Motion and fitness days are excluded.

To replace its recorded face, use Use this face for my Journal Soul.

Your face and voice, for family

Face and voice inclusion starts off. A package can include them only when their recorded permission allows sharing and use after death.

1. Open My Journal Soul, face & voice and, under My voice, for my family, tap Record my voice for my family.
2. Say the permission displayed, then record clear speech.
3. Review the separate permissions for receiving the recording and recreating the voice after death.
4. When preparing a family package, select My recorded face or My recorded voice only as intended.

Record my voice for this Soul can use the same saved family recording to create the Journal Soul’s voice through your own ElevenLabs account. Read that provider’s cost before confirming.

The recipient decides whether to keep a received face or voice. Free local face animation is available; recreating a voice requires their own eligible ElevenLabs account. Packages contain neither your API keys nor paid-provider account access. The result is an AI representation, not the person.

Memories left for you

1. Open Settings › Life Journal › Memories left for you. It is available before Life Journal is turned on; keeping received memories requires the journal.
2. Tap Turn on receiving for this account.
3. Save the receiving key’s recovery file somewhere private. It is needed to open received memories on a new iPhone.
4. Choose the email notification option if you want an email when memories are left for you. Emails do not contain the memories.
5. When a package is released, tap Open what they left you and read the message.
6. Choose whether to keep the recorded face and voice, then tap Keep these memories.

Received material stays separate from your own timeline under Kept in your journal. A .rbjournal file saved from reverbot.ai/life-journal opens only for the account it was released to; open it in Reverbot or use Import a package you saved from the website.

Hear their story, and bring in their Soul

In what someone left you, Their life story lists any chapters they wrote. Tap play to hear one told paragraph by paragraph, showing the photo, or playing the video without its sound, of each memory it cites; tap Read it myself to turn the pages instead. Write their life story asks the AI you choose to write one from the memories they approved — or all they left you when they approved none — citing each; it uses your own key or one of your free answers, and never sends their Health days. Under Their Soul, Bring them into Souls adds their Journal Soul — or makes one from what they left — with their photos, videos, recordings and kept face. It speaks as them, from those memories only, and it’s an AI: Reverbot says so at the start of every talk. Recreate their voice gives it the voice from their recording, on your own ElevenLabs account. That Soul’s own page lists their life story too, to read or hear in its voice. Your own life story drafts have the same play button, and are listed on your Journal Soul’s page.

If someone has died

Only a recipient the person chose can start the release process. Inactivity alone never releases memories.

1. Open Memories left for you and tap They have died — report it beside that person’s package.
2. Enter your name, the date and the requested details.
3. Add a death-certificate photo, PDF or a link to an official record or obituary.
4. Confirm the information is true to the best of your knowledge and submit the report.

Two independent reviewers must confirm against an independent source. The account owner is notified so a mistake can be stopped, and the waiting period they selected must pass before release. You can use Withdraw my report while review is pending.

Only reviewers see the record. It is deleted when the report closes, or 30 days after confirmation; reporter details are removed according to the stated retention period. The website offers this process for an email-and-password sign-in too.

If it doesn’t work

• If Record a memory is unavailable, turn on Enable Life Journal, sign in and stop any active listening session or call.
• If a recording cannot save, enter a Memory title and finish the recording first. Stop recording alone does not save it.
• If a source cannot be selected for a draft, review it and turn on Reviewed and approved for drafts.
• If a source permission is denied, change it in iPhone Settings; refreshing cannot override denial.
• If the soul-retention choice is unavailable, turn on Souls and create a soul first. The choice applies to saved interviews, not ordinary soul conversations.
• If a backup conflicts with another phone, read the restore-or-replace choices before proceeding. Do not overwrite a backup just to dismiss the message.
• If family memories cannot open on a new phone, restore the receiving key’s recovery file for the recipient account. A journal backup recovery kit and a receiving key are different files.

Still not working? In Settings, under Feedback, tap Contact Support. Never attach recovery keys or private journal material.

Before you begin

• Connect headphones and select them as this iPhone’s audio output. Bluetooth pairing alone does not select the playback route.
• Place the phone where its microphone can hear both people. Flirt uses Room, even though suggestions play in your headphones.
• Allow microphone and speech-recognition access.
• Have an internet connection and available AI questions or your own configured key.
• Use Rehearse first in Flirt settings to try the coach before a real conversation.

Configure your coaching

1. Touch and hold Flirt on the main screen and choose Settings.
2. Open About you and enter information you want the coach to use.
3. Open Goals, pacing and tone. Choose your Goal, Pacing and Tone.
4. In Coaching Style, choose Exact words to repeat for a proposed line or Coaching for guidance you can phrase yourself.
5. Review Silence duration before answering and Voice. Leave enough pause time for the other person to finish naturally.

Start a conversation

1. On the main screen, tap Flirt.
2. Tap the big button.
3. Let the other person speak first so Reverbot can follow the turns.
4. Wait for a suggestion when they finish. It appears under “SAY THIS” and plays in your headphones.
5. Reply in your own voice, using the suggestion only if it fits.

Flirt follows the next turns automatically. There is no finished-talking button to tap after every sentence.

An example to try

If they say “I moved here recently and haven’t explored much,” wait for the suggestion before replying. With Exact words to repeat, use the proposed line only if it fits you. With Coaching, use the idea in your own words. Keep listening to the real person; a suggestion is optional and the AI can misunderstand tone or context.

Skip, stop or start fresh

Tap the Flirt card while a suggestion is playing to skip it, or use an enabled skip gesture. End Conversation stops the session. Returning shortly can preserve the current conversation; use New on the conversation screen when speaking with someone different. This avoids carrying one person’s context into another conversation.

Use Flirt on a call

Important: The person called hears an automated AI announcement and must press 1 before Reverbot listens. You confirm twice before the call.

1. Tap the phone button, then Phone Call.
2. Open Call a number and choose Both of you.
3. Choose Flirt, then check the recipient and calling price.
4. Complete both warnings and wait for the recipient’s agreement.

The mandatory announcement applies in every US state. Starting Flirt on the main screen does not connect it to a separate Phone or FaceTime call. See the Phone calls topic for the complete flow.

If it doesn’t work

• If the wrong person’s turns are followed, touch and hold Flirt, tap Settings › Goals, pacing and tone, then tap Following the wrong person — start over. Let the other person begin.
• If speech is missed, uncover the phone’s microphones and place it where both people can be heard.
• If suggestions arrive too early, increase How long it waits under Silence duration before answering.
• If a headphones message appears, select the headphones as the active output and restart after the route settles. A room speaker is not a private coaching output.
• If suggestions concern the previous person, use New to begin a fresh conversation.

Still not working? In Settings, under Feedback, tap Report a Bug and include the active audio route.

Create a scenario

1. Touch and hold Custom Conversation on the main screen, then tap Settings.
2. Tap Your scenarios and sources.
3. Tap the Add button (+), enter a name and tap Add.
4. Open the template and fill in Scenario, The Other Person, You and Goal.
5. Return to Your scenarios and sources and tap the circle beside the template to make it active.

Only one template is active at a time. Changes in its editor save automatically; there is no separate Save button.

Create an Information Source

1. Open a template in Your scenarios and sources.
2. Under Purpose, choose Information Source in Use this template for.
3. Describe the subject and how you want answers explained.
4. Choose Only my materials or Materials + general AI knowledge.
5. Under Reference Documents, tap Upload reference files or Write or paste information. Add the facts, definitions and procedures the answers should use.
6. Wait for imports to finish and review any partial or failed files.
7. Activate the template, return to the main screen, choose Custom Conversation and tap the big button.

Ask about the material, for example “What does the handbook say about expense receipts?” An Information Source answers you; it does not play the other person or start a call.

Control what answers may use

Only my materials returns verified excerpts from the saved reference text, with document names. When the materials do not support an answer, it says so. Put factual definitions and policies in the reference materials, not only in the instructions field. Materials + general AI knowledge also permits the model’s existing knowledge, which can be incomplete or outdated. Neither option browses the web. Questions and relevant excerpts are still sent to the selected AI, so this is not an offline mode.

Use the source with Questions

1. In the source editor, tap Use with Questions. Alternatively, touch and hold Questions, tap Settings and choose the source under Answer from your information.
2. Ask your question while Reverbot is listening. The source’s answer policy applies, including when Questions runs beside translation.
3. Choose None — general questions to return to ordinary answers.

With Room, nearby people can ask too; Solo favors your voice. This can accompany a conversation on another device. iPhone may prevent access to the microphone while another calling app owns it.

On a phone-number call placed through Reverbot, Just you uses the selected Questions source, or an active source selected through Custom Conversation before dialing. For Both of you, choose Information Source and complete the mandatory recipient announcement and permission flow first.

Give the coach useful information

For a job interview, describe the role and interview stage, enter the interviewer’s name and priorities, list your actual experience, and state a concrete goal such as “Explain my leadership experience and agree on the next step.” Use my profile includes your Settings › About you information; turn it off when that background is irrelevant. Do not ask the coach to invent credentials or personal facts.

Add references and choose an AI

In the scenario, use Upload reference files (PDF, Office files, images, text and other supported formats) for a résumé, job description or notes. Assign each document as About me, About them or Shared context so the AI can distinguish the sources. Review the scenario’s AI choice; System Default AI at the bottom of Your scenarios and sources sets the starting choice for new scenarios. Provider keys, access and costs still apply.

Run the real conversation

Choose the live-coaching style in the scenario. Select its main-screen pill and tap the big button. Use headphones selected as the audio output and keep the phone microphone where it can hear both people; Room is selected for this mode. Listen to the suggestion, then decide what to say. Stop with the main listening control when finished. For private material, check that the active scenario and other person’s name are correct before starting.

Rehearse before using it live

Use the scenario’s Practice style or Rehearse first in Custom Conversation settings. Reverbot plays the other person and you answer aloud, with no real participant present. Live coaching and Standard practice listen in English; use a configured Live rehearsal partner for other supported conversation languages. End rehearsal before expecting the app to listen to a real conversation.

Choose the pill and button style

• Main-screen button name is optional. Leave it blank to show Custom Conversation; the field limits its length so the pill fits.
• The pill stays on the third Listening mode row beside Voice Changer.
• In the template editor, Button style gives that saved template its own look. Same as Custom Conversation uses the mode’s style.
• Saved conversations files conversations under the other person’s name. Open one to read it; touch and hold an older run to resume it.

Rehearsals start fresh and do not continue a real person’s conversation.

Export deliberately

Export template includes the scenario and extracted document text, not the original PDFs. Export with saved conversations additionally includes the saved conversation history. To bring a file back, open Your scenarios and sources › Import a conversation file; it creates a new editable template. Check sensitive attachments and history before sharing either export.

If it doesn’t work

• If the feature will not start, activate one usable template. A coaching scenario needs a name and scenario, other-person or goal information.
• If a source finds no answer, check its imported text. With Only my materials, unsupported answers are deliberately refused.
• If advice is generic, add a specific situation, actual background and a clear goal.
• If it refers to the wrong person, check Their name and the selected saved conversation.
• If answers fail, check the template’s AI and its key or allowance.
• If you expected an exact transcript, review Keep their exact words; rehearsal transcripts are kept in full.

Still not working? In Settings, under Feedback, tap Report a Bug. Do not attach private documents unless needed and permitted.

Before you begin

• Tell the participants you intend to record and explain how you will use the meeting. Follow Reverbot’s announcement and agreement flow.
• Allow microphone and speech-recognition access, and position the phone where the room is audible. AI Meeting uses Room.
• Touch and hold AI Meeting, then tap Settings. Review Keep the audio recording, Report and Keep meetings for.
• Leave Analyse automatically when a meeting ends and Email the report to participants automatically off if you want to decide after each meeting. Review Include the transcript in that email separately.

Start an agreed recording

1. Tap AI Meeting and the big button, or open AI MEETING and tap Start the meeting.
2. If prompted, say “start the meeting” so Reverbot can establish the organizer’s voice.
3. Let the recording notice finish.
4. Wait for the required agreement from someone else and the “Recording” status.

Only then has recording begun. A participant joining from their own phone reviews and agrees there before contributing. Tapping Start alone does not authorize a saved recording.

Use the meeting screen

Open AI MEETING to see This meeting, the participants and transcript. Mark this moment adds a bookmark for something you want to find later. People joining from their own phones review the request and agree there; their phones contribute their own side. Use the room’s message and contact-card controls only for information you intend to share with those participants.

End normally or withdraw consent

End the meeting finishes an agreed recording for later review. An objection is different: “stop recording” or a participant’s Withdraw revokes agreement and triggers deletion under the meeting flow. If Reverbot is speaking, a spoken objection may not be recognized then; repeat it after speech ends, use Withdraw on the participant’s phone, or have the organizer stop immediately. The notice and agreement requirement cannot be disabled.

Generate and review the report

1. Open AI MEETING › Past meetings, or Meetings in AI Meeting settings.
2. Open the completed meeting and review its transcript.
3. Check the displayed analysis estimate, then tap Analyse now. Analysis sends the relevant transcript to the selected AI.
4. Read the report against the transcript, especially names, dates and action items.
5. Use Share the report, Share as PDF or the transcript/audio sharing controls as needed. Very long transcripts can be limited; read the analysis explanation before relying on complete coverage.

Plan ahead or use video

Plan a meeting in AI Meeting settings stores the planned details and invitation choices. A planned meeting is not itself consent to a future recording. Start a video meeting uses the configured video/Cloudflare setup and invites up to 7 friends; each participant must agree to recording before it starts. A phone-number AI Meeting uses the Phone Call › Both of you warnings and recipient permission flow as well.

Optional answers and translations

Answer questions too and Translate too are available in the meeting settings and on the Meeting screen. They can produce audible speech, including speech from the phone for the room. Read their separate guides before enabling them. While Reverbot speaks, the speech transcript can miss overlapping room speech even when the audio recording continues.

If it doesn’t work

• If the status is waiting for agreement, let the announcement finish and obtain the requested participant agreement. It is not a microphone failure.
• If no words appear after “Recording”, check microphone and speech-recognition permissions and phone placement.
• If analysis is unavailable, check whether the transcript contains any speech and whether the selected AI has a key or allowance.
• If a report seems incomplete, compare it with the transcript and read any length limit shown before analysis.
• If someone withdraws while Reverbot speaks, use Withdraw on their joined phone or have the organizer stop. Spoken objections may not be heard during playback.

Still not working? Keep the meeting if consent still permits it, resolve the displayed error and retry analysis. Use Settings, then Report a Bug under Feedback, for a persistent problem.

Where it works

• It works beside AI Meeting, Custom Conversation, Translate Vicinity and Translate 2-way.
• It isn’t offered for Questions or Dictionary, which answer questions themselves, or for an Information Source. Flirt, Job, Personal Assistant and Home Control don’t offer it either.
• It doesn’t answer during a rehearsal, or on a phone call with Both of you, where the person you called agreed only to the feature you chose. On a Just you call, Questions beside Translate works as it does in person.
• It needs an internet connection, and free questions or your own AI key.
• In AI Meeting, answers start only once the meeting is recording.

Turn it on with a translation

1. On the main screen, tap Translate Vicinity or Translate 2-way.
2. Tap Questions. Both pills light up.
3. Tap the big button.

To turn the answers off, tap Questions again; the translation keeps running. The same switch is Answer questions too in Vicinity Language Translation and Two-way Language Translation settings. Both translation modes share it, so it stays on when you swap between them.

Turn it on in other modes

1. Touch and hold AI Meeting, or the Custom Conversation pill, then tap Settings.
2. Turn on Answer questions too.
3. Tap Done, then start the meeting or conversation as usual.

AI Meeting and Custom Conversation share this one switch. It’s also on the Meeting screen, which AI MEETING in the top bar opens: under Questions and translation, turn it on or off before or during a meeting.

What happens when someone asks

For example, during a meeting someone asks “What’s the difference between a fixed and a variable cost?” When they stop, Reverbot answers as soon as it isn’t speaking itself, and the meeting keeps recording.

• Beside a translation, a question in another language is answered after its translation has been read to you.
• In AI Meeting, the answer goes into the transcript as its own line, marked Reverbot, so the report includes it.
• Statements, small talk and opinions get no answer, as in Questions.
• Answers play in your headphones. Without headphones, they’re spoken only where Reverbot already speaks out loud, as in Translate 2-way, or when Speakerphone is on in Settings; otherwise they appear on screen.

What it costs

• Each answer uses one of your free questions, like any Questions answer, or a request on your own AI key. The AI chosen under AI in Questions settings answers these questions too.
• A question that gets no answer costs nothing.
• When your free questions run out, the answers stop and the main mode carries straight on; nothing interrupts it. After adding a key, turn the switch off and on to start answering again.
• While Reverbot is speaking, it can’t hear the room, so a few seconds of speech can be missing from a meeting’s transcript. The recording keeps them.

Objections in a meeting

Important: Reverbot doesn’t hear an objection while it’s speaking. Anyone who doesn’t want to be recorded should say it again once it stops, or tap Withdraw on their own phone if they joined from it.

Answering questions never changes a meeting’s announcement or the agreement it needs. The AI Meeting topic explains objections and how to end a meeting.

If it doesn’t work

• If nothing is answered, check that the switch is on for the mode you’re using: the translation modes share the Questions pill, and AI Meeting and Custom Conversation share the switch in their settings.
• If answers stopped partway through, your free questions may have run out. Add your own AI key, then turn the switch off and on.
• If answers come too early or too late, change How long it waits in Questions settings: Answer questions too uses the Questions timing.
• If nothing is answered in a meeting, check that it’s recording. Answers start only after everyone has agreed.
• If answers appear on screen but aren’t spoken, connect headphones or turn on Speakerphone in Settings.

Still not working? In Settings, under Feedback, tap Report a Bug and describe what you did, what you expected and what happened.

Before you begin

• Translate too works only with AI Meeting. Translate Vicinity and Translate 2-way already translate, and other modes don’t offer it.
• It doesn’t run on a phone call or during a rehearsal.
• Your language is Translate to in Vicinity Language Translation settings. Nobody chooses the room’s language: Reverbot detects it from what’s said.
• The meeting’s announcement and agreement come first, as always. Translation starts once the meeting is recording.
• Have an internet connection, and translation available: free questions, a subscription or your own Gemini or OpenAI key.

Turn on Translate too

1. Touch and hold AI Meeting, then tap Settings.
2. Turn on Translate too.
3. Check Out loud on speakerphone, under it.
4. Tap Done.

You can also turn it on or off on the Meeting screen, which AI MEETING in the top bar opens: it’s under Questions and translation, before or during a meeting. If people have joined from their own phones when the meeting starts, Translate too comes on for that meeting by itself.

One phone for the whole room

When nobody has joined from their own phone:

• When someone else speaks, their sentence is translated into your language and plays in your ear.
• When you speak, your sentence is translated into their language and plays from the iPhone’s speaker, even with headphones in.
• Reverbot learns the room’s language from the first sentence it translates for you, so let someone else speak first. Keep the iPhone close to you, so it can tell your voice from the room’s.
• Pause while a translation plays. Reverbot can’t hear the room then, though the recording keeps those seconds.

In the transcript, their words appear in your language, marked “them”. Your own words are already there, so the version said for the room isn’t added twice.

Control the room speaker

Out loud on speakerphone sends your translated sentences out of the iPhone’s speaker, even with headphones in. Turn it off to keep them in your ears.

• While one phone translates for the room, a Speaker pill appears under Input & answer on the main screen, and the switch appears on the Meeting screen too. Tap Speaker to switch to Earpiece.
• In AI Meeting settings, the switch can be changed only while Translate too is on.
• It’s the same setting as Translate 2-way’s, and it turns back on each time a meeting starts.

When people join from their phones

If the people with you joined the meeting from their own phones, each of them hears every translation on their own phone, in their own language, and nothing is said out loud. What they say comes back to you in your language.

Once they’ve agreed to the recording, it goes into the transcript under their name, so the report covers the whole room. The AI Meeting topic explains how people join.

Settings that change it

• Translate to, in Vicinity Language Translation settings, is your language.
• How long it waits, under Silence duration before answering in the same settings, sets how long a pause ends a sentence for Translate too.
• AI, in the same settings, chooses who translates: Reverbot (included), or Gemini or ChatGPT on your own key.
• Answer questions too, beside Translate too, answers questions as well: a question in another language is translated first, then answered.
• Restore AI Meeting defaults turns Translate too off and Out loud on speakerphone back on, with the rest of AI Meeting’s settings, after you confirm.

What it costs

• Before you subscribe, each translation uses one of your free questions. With a subscription, translations come from a daily allowance instead.
• With your own Gemini or OpenAI key chosen under AI in Vicinity Language Translation settings, translation runs on your key, and that company bills you.
• Anything already in the right language costs nothing.
• If translation runs out, Translate too stops and the meeting carries on; nothing is interrupted. Turn it off and on again later to try again.

Objections and stopping

Turn off Translate too to stop translating; the meeting keeps recording.

Important: Reverbot doesn’t hear an objection while it’s speaking a translation. Anyone who doesn’t want to be recorded should say it again once it stops, or tap Withdraw on their own phone if they joined from it.

The AI Meeting topic explains objections and how to end a meeting.

If it doesn’t work

• If nothing is translated, check that the meeting is recording: translation starts only after everyone has agreed.
• If you hear nothing in your language, check Translate to and whether people are already speaking your language.
• If your sentences aren’t translated for the room, let someone else speak first so Reverbot learns their language, and keep the iPhone close to you.
• If someone else’s words are treated as yours, keep the iPhone nearer to you than to them, and take clear turns.
• If the room can’t hear your translations, check that Out loud on speakerphone is on and the volume is up.
• If translations stop partway through, translation may have run out or the connection dropped. Check the connection, then turn Translate too off and on.

Note: Translations can be wrong. Check names, numbers and commitments with the speaker.

Still not working? In Settings, under Feedback, tap Report a Bug and describe what you did, what you expected and what happened.

Prepare the conversation

1. Touch and hold Flirt, Custom Conversation, Translate Vicinity or Translate 2-way, then tap Settings.
2. For Custom Conversation, create and activate a scenario describing the situation and the person you want to practice with. Flirt uses its coaching context.
3. Under Practice, turn on Rehearse first.
4. Choose Partner’s voice and any available language setting.

Finish a current session before changing rehearsal options. Practice cannot start during a connected call.

Turn on Rehearse first

Find Practice and enable Rehearse first. This changes the selected mode into a rehearsal when you next start it. Changing rehearsal options can end a session that is already running, so finish your current conversation before setting these options.

Choose the partner’s voice

Important: A live rehearsal streams microphone audio to its voice provider. Read the selected option’s access and cost information before starting.

• Standard takes turns: speak, pause and hear the response. Choose Standard partner’s AI for its text responses.
• Live — Reverbot uses Reverbot’s available live-voice allowance.
• Live — OpenAI (your key) uses your own OpenAI API key and provider billing.

Translate Vicinity uses the Standard partner even if a live option is selected. For other supported modes, the selected partner determines interruption and turn-taking behavior.

Set a practice language

In either translation mode, use Partner speaks to choose the language you want to hear. For example, choose Spanish to practice ordering a meal. Translate Vicinity always uses the Standard partner, even if a live option is selected. A Custom Conversation uses its scenario language.

Start and take a turn

Return to the main screen and tap the big button. Listen to the partner, then answer naturally. With Standard, leave a pause after your turn. With a live partner, you can speak over the response to interrupt it. Use the transcript to check what Reverbot understood.

End Practice and return to the real mode

Tap the big button to end the rehearsal. Before using the mode with another person, return to its Settings and turn off Rehearse first. Rehearsal transcripts are saved in full, even if the scenario normally keeps only a summary of the other person’s words.

If it doesn’t work

• If a rehearsal will not start, end any connected call and check the internet connection and selected partner’s key or allowance.
• If it coaches you instead of playing the other person, turn on Rehearse first before starting again.
• If the wrong language is used, review Partner speaks, or the Custom Conversation scenario language. Standard coaching and ordinary Standard practice listen in English; use a supported live partner for other conversation languages.
• If the partner interrupts too soon, adjust Silence duration before answering and read its explanation for that partner.
• If you are ready for the real conversation, stop with the big button, then turn off Rehearse first.

Note: Rehearsal transcripts are saved in full, even when the scenario normally keeps only summaries of the other person’s words.

Turn on the wake phrase

1. Open Settings.
2. Under Wake Phrases, turn on Job.
3. Read the phrase below it; the default is “hey job”.
4. Tap Job settings to configure workspaces, voice and timing.

Turning on a wake phrase does not start the microphone. Start listening with the big button before saying it.

Create a workspace

1. In Settings, under Wake Phrases, tap Job settings.
2. Tap Jobs & Workspaces.
3. Tap the Add button (+), enter a Job name and tap Add.
4. Open the workspace.

Its Tasks, Projects, Calendar, Reminders, Lists, Notes, Docs and Setup tabs hold the work for that job. Scroll the tab row sideways to see the remaining tabs. Use separate workspaces for separate jobs.

Choose where commands go

In Jobs & Workspaces, swipe a job row and tap Default to make it the default work destination. Include a job’s name when you want a different destination, such as “hey job, add Acme task prepare the estimate”. Check the saved item in that job after your first command.

Start listening and speak one command

Return to the main screen and tap the big button if Reverbot is inactive. Say the wake phrase and the instruction together: “hey job, add task call the bank”, “hey job, remind me to submit the invoice tomorrow at 9 AM”, or “hey job, take a note: the delivery entrance is on Oak Street”.

Confirm the result or undo it

Listen to the confirmation and check the item’s title, date and destination when needed. If Reverbot asks a follow-up, answer it. After a supported change, say “undo”, “never mind” or “cancel that” within the undo window to reverse it. A new command closes that window; the one-shot command returns to your listening session when finished.

Draft messages and manage lists

Try “hey job, what are my tasks” or “hey job, add to grocery eggs”. Text and email commands prepare a draft for you to review and send yourself. “Buy paper towels” adds a Shopping item and can offer a store link; it does not place an order or pay for anything.

Customize the phrase and timing

Open Job settings › Wake phrase & behaviour. Edit Job wake phrase; an empty field uses “hey job”. Undo Window can be set from 3 to 15 seconds. Silence After Trigger controls how long an unfinished command waits for a prompt. To change command wording for one job, open that workspace’s Setup tab.

If it doesn’t work

• If a command is ignored, check that Job is on under Wake Phrases, the big button is listening and the phrase matches the one shown.
• If the wrong workspace is used, name the job in your command and check the Default job in Jobs & Workspaces.
• If a saved item has the wrong time, use a full date and time, then review the result.
• If a command is misunderstood repeatedly, review that workspace’s Setup bindings and say one instruction at a time.
• If no copy appears in Calendar or Reminders, turn on the corresponding option under Calendar & Reminders Sync and select a writable destination.

Still not working? In Settings, under Feedback, tap Report a Bug and include the command without private information.

Turn on Assistant

1. Open Settings.
2. Under Wake Phrases, turn on Assistant.
3. Read the phrase shown below it; the default is “hey assistant”.
4. Return to the main screen and tap the big button.

Job and Assistant can be on together. Their wake phrases work within an active listening session.

Give a specific instruction

Say “hey assistant, remind me to call Mom tomorrow at 5 PM”, “hey assistant, add to grocery eggs and milk”, or “hey assistant, take a note: the spare key is in the blue drawer”. Include a date and time for timed reminders. Answer any follow-up question, then listen for confirmation.

Review your personal workspace

1. Open Settings.
2. Under Wake Phrases, tap Assistant settings.
3. Tap Jobs & Workspaces and open the workspace under Personal Assistant.
4. Choose Tasks, Calendar, Reminders, Lists or Notes to review and edit the result.

These are your personal items. Work jobs have separate workspaces.

Correct or finish a command

Say “undo”, “cancel that” or “never mind” during the undo window after a supported change. To abandon an unfinished command, say “cancel”. When the command finishes, Reverbot returns to the listening session underneath it. Tap the big button to stop listening altogether.

Send an item to work deliberately

To use your default job instead of the personal workspace, include “work”, for example “hey assistant, add work task finish the estimate”. To target a specific job, use its name: “hey assistant, add Acme task call the supplier”. Set the default job in Jobs & Workspaces.

Assign a soul as assistant

1. Open SOULS, choose the soul and tap Permissions.
2. Turn on Make [name] your assistant, for example Make Sally your assistant.
3. Review the organizer access this role includes and any optional permissions.
4. In Settings, under Wake Phrases, keep Assistant on.

The assistant phrase now calls that soul instead of the built-in assistant. Only one soul can hold the role. A soul must live in the phone and know it is an AI; a soul with a life of its own or configured to believe it is alive cannot control the phone or be your assistant.

Customize it or turn it off

Open Assistant settings › Wake phrase & behaviour to change Personal Assistant wake phrase, the undo window and silence timeout. Leaving the phrase empty restores “hey assistant”. To return to the built-in assistant, turn off the assistant role in the Soul’s Permissions. To ignore the phrase entirely, turn off Assistant in Settings, under Wake Phrases.

Check reminders and external actions

Review the saved time and allow notification access if you want alerts. Copying items into iPhone Calendar or Reminders is a separate opt-in setting. Texts and emails still require you to review and send the draft; shopping links take you to the store to complete any purchase.

If it doesn’t work

• If nothing happens, turn on Assistant under Wake Phrases and start listening with the big button.
• If the wrong assistant answers, review the assigned soul’s Permissions. Turn its assistant role off to return to the built-in assistant.
• If an item goes to work unexpectedly, omit “work” and job names for personal items.
• If a reminder does not alert you, check its saved time and Reverbot’s notification permission in iPhone Settings.
• If a draft is not sent, open and review it, then send it yourself. Reverbot does not silently send messages.

For Calendar and Reminders copying, see the Calendar & Reminders topic.

Availability comes first

The public app shows Coming soon while Reverbot’s Google Home connection awaits approval. Turning on a wake phrase or saving a location does not make device control available. Check the status in Home Control settings after updating the app; there is no setting that bypasses this requirement.

Find its availability

1. Open Settings.
2. Under Wake Phrases, tap Home Control settings. Its name may follow your selected location.
3. Tap Google Home & locations.
4. Read Availability before changing device permissions.

“Coming soon” means device control is not available in this release. Saving a location or turning on a wake phrase does not change that.

Prepare location names

In Locations, tap Add a location or edit a saved location, enter its Name and choose Home or Office. Tap Save. Tap a saved location to make it active. The default wake phrase follows that kind: “hey home” for Home or “hey office” for Office. An optional Short name is limited to 14 characters.

Review the wake phrase

Open Home Control settings, then Wake phrase & behaviour. A custom Home Control phrase applies to every location; leaving it empty uses the active location’s default. Keep the switch off if you do not want that phrase handled while you listen. There is no wake-phrase section on the main screen.

What the planned commands are for

The intended workflow is to choose an authorized Google Home location, then name a device or room in a command, such as “hey home, turn off the kitchen lights”. These examples describe the feature being prepared; they are not a promise that the current public release can control lights, locks or thermostats.

If it doesn’t work

• If commands do nothing, check whether Home Control still says “Coming soon”. No setting bypasses this release restriction.
• If a location name is unexpected, review Locations and its Short name.
• If the availability message seems wrong after an app update, open Settings, find Version, then contact support under Feedback, using Contact Support.

Use your existing home-control app for devices while this feature is unavailable.

Choose the listening feature first

Pick Questions, Dictionary or another listening-mode pill. Some features set their own microphone because they need to hear another person. Make this choice before adjusting Room or Solo, so you can see whether the feature allows a microphone preference.

Use Room for surrounding speech

In Input & answer, tap the microphone pill to select Room where the choice is available. Room uses the phone microphone to pick up nearby speech. Put the phone close enough to the people you want to hear and keep its microphones uncovered. Headphones can still receive the spoken answers.

Use Solo for your own questions

Select Solo when you want the microphone setup to favor your voice. With a compatible headset, this can use its microphone; otherwise the available phone route is used. Solo is a microphone preference, not a guarantee that other voices are excluded and not voice identity verification.

When the microphone choice is locked

Flirt, both translation modes, Custom Conversation, AI Meeting and rehearsals manage the microphone themselves. Room is used for surrounding conversation; a meeting joined through participants’ phones may select Solo for each participant. Your previous preference returns when the mode releases it. A locked pill does not mean the microphone has stopped.

Select Concise or Extended

Tap the answer-length pill to switch between Concise and Extended. Concise requests a short, direct answer; Extended asks for more explanation. The selected AI and feature determine the final wording. This control is not shown for Flirt, and it does not change speech speed or the pause before answering.

Keep microphone and output choices separate

Room and Solo choose how Reverbot listens. Speakerphone in Settings controls whether ordinary answers may play out loud without headphones. With it off, those answers are shown on screen when no private output is available. Features designed to speak to someone, such as two-way translation and rehearsals, have different output needs.

If it doesn’t work

• If Room or Solo will not switch, check the selected mode. Conversation, translation, meeting and rehearsal features can manage their own microphone.
• If no words appear, allow Microphone and Speech Recognition for Reverbot in iPhone Settings, then restart listening.
• If the microphone changed after a headset connection, stop listening, choose the intended input under Microphone Input in Settings and start again.
• If an answer is visible but silent, check the headphone output or Speakerphone; microphone choice does not select where the answer plays.
• If speech is too fast, change Voice speed in Settings › Voice & Speed, not Concise or Extended.

Start the session before pocketing the phone

Choose the listening mode, tap the big button and confirm that the status says it is listening. Make sure the audio is routed where you expect. You can then lock the screen. A wake phrase responds within an active listening session; it does not start the microphone from a fully stopped app.

Skip with the headset’s Next Track gesture

While Reverbot is speaking, use the gesture your headphones assign to Next Track. Reverbot treats that media command as skip. The exact press or tap depends on the headset and its configuration, so use the gesture shown in your device’s settings. This ends the current answer; it does not ask the question again.

Turn on Shake to Skip

1. Open Settings.
2. Under Skip Answer, turn on Shake to Skip.
3. While an answer plays, shake the phone to end that answer.

Turn the switch off if ordinary movement skips answers accidentally. Tap to skip on screen remains available.

Stop from the active media controls

While Reverbot owns the active media session, the headset or lock-screen play/pause control toggles the session. Confirm the stopped state when practical. After a full stop, Reverbot releases those media controls so they can belong to another app; open Reverbot and use the big button to start again.

Replay an answer

When Repeat that is available below the big button, tap it to hear the last answer again. It replays the existing response rather than submitting a new question. Check that the intended headphones are the active output if an answer appears on screen but is not audible.

Choose audio cues

Open Settings and find Audio cues. Turn it on for the short sounds used for actions such as listening, completion, undo and cancellation. The Audio cues switch in Wake phrase & behaviour changes the same preference.

Turning cues off removes those sounds, not the spoken answers.

If it doesn’t work

• If a headset gesture controls music, check that Reverbot still has an active listening session and another audio app or call has not taken the media controls.
• If skipping fails, try Tap to skip on screen, then use the headset gesture assigned to Next Track. Headset gestures vary by model and configuration.
• If a wake phrase does nothing with the screen locked, reopen Reverbot and confirm that the big button started a listening session. A wake phrase cannot start a stopped microphone.
• If an interruption ends the session, return to the app and start listening again.

The iPhone’s volume, side and Action buttons are not Reverbot skip controls.

Choose and hear a voice

1. Open Settings › Voice & Speed.
2. Under Voice Engine, choose Reverbot Voice, On-device (Apple) or a provider for which you have your own key.
3. Choose a voice using the controls shown for that engine.
4. Tap Test voice.
5. Adjust Volume and your iPhone’s playback volume as needed.

For a cloud engine, On-device voice chooses the fallback voice, not its normal online voice.

Change the mode’s voice quickly

1. While Reverbot is inactive, tap the sliders button (three short lines with knobs) under the big button.
2. In Button & Voice, tap Voice at the top.
3. Tap a voice’s play button to hear a sample. A preview does not choose it.
4. Tap the voice to use it.
5. Drag Speed to adjust delivery, then tap Done.

The page shows the voice the current mode uses. Answers produced by Questions beside translation still use Questions’ voice.

Choose what the change affects

The Voice page in Button & Voice ends with Only for [mode], for example Only for Questions.

• Off: your choice changes the default used by modes without their own voice.
• On: the choice applies to this mode.
• Turning it on after choosing keeps the choice for this mode and restores the default this page changed.

In a saved Custom Conversation, the voice switch is Only for Custom Conversation; individual scenarios have their own button styles, not separate copies of this voice switch.

Note: With Coaching cues on, Flirt and live Custom Conversation can whisper using a configured OpenAI, Gemini or Grok voice. Choosing a voice may give that mode its own override so you actually hear the choice.

Review voice costs

In Button & Voice › Voice, services are grouped by access:

• Ready to use includes Reverbot Voice and the iPhone’s installed voices.
• Better quality, free to start includes services with an available provider allowance. Check the service page for eligibility and billing after that allowance.
• Paid from the start identifies services billed from first use.

Important: A voice on your own key is billed by that provider, separately from Reverbot. Read the current price and account requirements in its setup page. A subscription to the provider’s chat app is not necessarily API or voice access.

Add a provider voice

Important: The selected company bills your own account. Google Cloud may require a billing account even when usage fits a free allowance.

1. In Button & Voice, tap Voice, then a service such as Gemini voices.
2. Read the allowance and price box.
3. Use its key-help link if you need a key.
4. Paste the key in its field and tap Save the key.
5. Optional: Tap the play button beside a voice to preview it.
6. Tap a voice to use it.

The app tries the voice with your key before choosing it. If the request fails, read the error. Some Google voice samples can be heard before adding a key.

Adjust speech speed

Drag Voice speed in Settings › Voice & Speed between 0.5× and 2×, then tap Test voice.

Speaking speed changes how quickly the answer is read. Silence duration before answering changes how long Reverbot waits for you to finish; it is a different setting.

The on-device Pitch control applies only where that engine supports it.

Give a mode its own voice

1. Open Settings › Voice & Speed.
2. Open Voice for each mode or Speed for each mode.
3. Turn on the override for the feature and choose its voice or speed.

The same controls appear in that feature’s settings. A soul has its own AI & voice page and uses its selected voice when it speaks.

Record a soul’s voice

Important: Voice cloning sends the sample to your own ElevenLabs account and requires a plan that supports it. Record yourself or someone who explicitly permits cloning and use.

1. Stop listening or any call.
2. Open the soul’s AI & voice page and tap Record a voice for this Soul.
3. Tap Record sample, record 10–120 seconds of clear speech, then tap Stop recording.
4. Tap Play sample, or use Record again. One to two minutes in a quiet room gives more material.
5. Enter a name and confirm that the voice is yours or that the speaker permits its use.
6. Tap Create voice and attach to [name].

If ElevenLabs requests verification, finish it in that account, refresh its voice list and choose the result.

Manage the recorded voice

The cloning sample goes directly to ElevenLabs. Manage or delete the resulting voice in that account; closing the recorder or deleting a soul does not delete the provider’s voice model. The temporary sample is removed when this recorder closes.

Life Journal’s separate family-voice recording has its own retention, permission and sharing controls. A generated voice creates new speech; it is not playback of the original person.

If it doesn’t work

• If the voice sounds different, read Last line spoken by at the top of Voice & Speed. A failed cloud request can use the iPhone voice; the reason is shown there.
• If a paid voice fails, check the provider key, enabled service, billing and access to that voice.
• If only one mode sounds different, check its override and Coaching cues.
• If an iPhone voice sounds basic, download an enhanced or Premium voice in iPhone Settings, then choose it in Reverbot.
• If the selected sound does not return, use Restore voice defaults and review the confirmation. API keys and button styles remain.

For on-device speech everywhere, also remove cloud overrides from modes and souls.

Open the sync controls

1. Open Settings.
2. Under Wake Phrases, tap Job settings or Assistant settings.
3. Find Calendar & Reminders Sync.

Both screens control the same mirroring preferences. Existing workspace items can be copied when you turn mirroring on, so choose your destination deliberately.

Grant the required access

Turn on Add events to Calendar or Add tasks & reminders to Reminders. Allow Full Access when iOS asks. Reverbot needs to locate its previously created copies so it can update or remove them; Calendar access limited to adding events is not enough for this workflow.

Select the destination

After enabling a switch, tap Calendar or List beneath it. Choose a writable destination; the picker includes its account source so you can distinguish iCloud, Exchange, Google or other available accounts. If you leave Default selected, the iPhone’s default destination is used.

Create a test item and verify it

Create an event or reminder in Reverbot with a clear title and date, such as “hey assistant, schedule planning time tomorrow at 10 AM”. Open the Calendar or Reminders app and find the copy in the selected destination. Review the time and account before relying on a notification or sharing the event.

Use an Outlook or other account

First add and turn on the account in the iPhone’s system settings for Calendar or Reminders. Return to Reverbot and choose the account’s writable calendar or list if iOS makes it available. Account support varies: an Outlook calendar appearing does not guarantee that the account also exposes a reminders list.

Understand the direction and scope

This is not a full import of your external calendar. Items flow from Reverbot into the selected system destinations; changes made only in another app are not imported back as Reverbot items. Turning mirroring off stops future copying and updates but does not remove copies already created. Review external copies separately when cleaning up.

Share a calendar file instead

Open a workspace’s Calendar, Tasks or Reminders tab. Touch and hold the item and choose Share as calendar file, then select a destination from the share sheet. The exported .ics file can be imported into a compatible calendar app. A file export is a one-time copy, not a continuing sync connection.

If it doesn’t work

• If items do not appear, check that the correct sync switch is on and the selected Calendar or List is writable.
• If permission was removed, open iPhone Settings and restore Reverbot’s Full Access to Calendar or Reminders, then turn mirroring on again.
• If a destination is missing, confirm its account is enabled in the iPhone’s Calendar or Reminders settings. An Outlook calendar does not guarantee a reminders list is available too.
• If edits made in another app do not return to Reverbot, that is expected: this is one-way copying from Reverbot.
• If copies remain after turning mirroring off, remove those external copies separately. Turning the setting off stops future copying and updates.

Review the feature before using it

Open Settings › Privacy Policy. Speech recognition can run on the phone where supported; this does not mean every feature is offline. Translation can send microphone audio for processing, live voice sessions stream audio, AI features send relevant text or files, and cloud voices send the words they will speak.

Control device permissions

Use iPhone Settings to review Reverbot’s microphone, Speech Recognition, Photos, Camera, Contacts, Calendar, Reminders and other requested permissions. Grant access for the features you intend to use. Revoking access can disable that feature; it does not automatically delete information you previously saved, exported or sent to a provider.

Review recording and sharing choices

AI Meeting and recorded video meetings have their own notice and agreement flows. Life Journal recordings and Soul voice samples begin through their recording controls. Check the feature’s save, analysis and sharing settings before starting; live processing and a saved recording are different data flows. Do not assume disabling one recording switch disables all processing.

Protect your provider keys

Reverbot stores provider API keys in the iPhone’s Keychain. Treat each key as account access: do not paste it into a support message, shared document or screenshot. Remove an unused key in its settings and revoke a compromised key with its provider. Choosing Apple’s on-device voice changes speech synthesis, not the separate AI-answer or translation service.

Manage saved data

Use each feature’s history, export and delete controls for saved items. To delete the Reverbot account, open Settings, scroll to the end and tap Delete account.

Important: Read the deletion confirmation and export anything you need first. Account deletion cannot retrieve copies shared with other people, and separate provider accounts may require deletion there. Deleting the account or app does not cancel an Apple subscription.

Ask for help or report a problem

Open Settings and, under Feedback, tap Contact Support, Report a Bug or Request a Feature. Describe the mode, the steps you took, what you expected and what happened. For audio problems, include the headset, selected input and whether the screen showed an answer. Tap Send and wait for the “Sent” confirmation.

Replies and submission details

The form adds the app version, iOS version and device information. Support replies use the email on your signed-in account; guest submissions have no reply address, so sign in first if you need a response. An unsuccessful submission shows an error—check the connection and retry, or use support@reverbot.ai. Avoid including another person’s private conversation unless it is necessary and you have permission.

If it doesn’t work

• If a permission-dependent feature stops working, review Reverbot’s permission in iPhone Settings. A refresh cannot override a denied permission.
• If a feedback submission fails, check the connection and retry. Use support@reverbot.ai if the form remains unavailable.
• If you need a reply, sign in first: guest feedback has no reply address.
• If a key may have been exposed, revoke it with the provider and replace it in Reverbot. Never include keys in support messages or screenshots.

In Settings, under Feedback, choose Contact Support, Report a Bug or Request a Feature. Describe the feature, your steps, the expected result and the actual result.

Find the right settings

Use the top Settings button for shared settings such as AI & API Keys, Voice & Speed, Microphone Input and Wake Phrases. For one listening feature, touch and hold its pill and choose Settings. Wake phrases have no main-screen section; their switches and configuration are in Settings, under Wake Phrases.

Verify capture with a simple phrase

Pick Questions or Dictionary, start listening and say a short, clear phrase, such as “What is the capital of France?” or “Define ephemeral”. Watch the transcript. If no words appear, check microphone and Speech Recognition permissions, the selected input and any displayed audio error before changing the AI model.

Check AI access when words appear but no answer follows

Open Settings › AI & API Keys and review the selected provider, key and available allowance. A provider-specific error can indicate a missing key, unavailable model or account limit. Try one simple request after correcting the cause. Repeatedly tapping start does not repair a provider account or restore a spent allowance.

Check playback when an answer is visible

Make sure your intended headphones are the active output, not merely paired in Bluetooth settings. Check device volume, Voice & Speed › Volume and Last line spoken by in Voice & Speed. With Speakerphone off, ordinary answers can be shown silently when headphones are unavailable. Turn it on only when you intend the answer to be heard aloud.

Adjust one timing setting at a time

If Reverbot interrupts before you finish, increase the relevant feature’s setting in Silence duration before answering. If it waits too long, reduce that setting and test a complete sentence. Voice speed is a separate control for playback. Restore a feature’s default before comparing large changes.

Stop, then change the setup

End the current session with the big button before swapping headsets, changing a provider key or making several mode changes. Confirm the status returns to “Inactive”, apply the change and start again. A rehearsal or room-dependent mode can manage its own microphone; a locked Room/Solo control is expected there.

Use the guide for the feature you are in

Each mode’s Settings includes How [feature] works. Use that guide for setup, examples and limitations, then return to the mode and try one step. If the same problem remains, Report a Bug under Feedback in Settings accepts a reproducible set of steps and adds device/build details automatically.

If it doesn’t work

• If there are no recognized words, troubleshoot permissions and the microphone before changing the AI.
• If words appear but there is no answer, check the feature’s request rules, key, connection and allowance.
• If an answer appears without sound, check the active headphone output, volume and Speakerphone.
• If the voice is unexpectedly robotic, open Settings › Voice & Speed and read Last line spoken by and any cloud-failure reason.
• If a mode behaves like a rehearsal, turn off Rehearse first before starting it with another person.

Still not working? In Settings, under Feedback, tap Report a Bug. Include the mode, what you said or tapped, the audio route and the visible message. Omit API keys and private conversations.

Before you begin

• You need an account with the company. A Gemini key needs a Google account, even if you use Reverbot with Apple.
• Most companies need a payment method or prepaid credit before a key works. Gemini is free within Google’s daily limits, which AI Studio shows.
• A ChatGPT, Claude or Grok subscription doesn’t include API use, and a password is never an API key.
• Copy the whole key with the company’s copy button. Some companies show a key only once.
• Keys are kept in this iPhone’s Keychain and sent only to their own company. Never put a key in a message or a screenshot.

Find your AI settings

Open Settings (the gear), then tap AI & API Keys. The AI Provider & Keys screen opens:

• At the top, Answering now names the AI answering you, and how: “On Reverbot’s connection”, “On your own key” or, for example, “On your Claude key, while Gemini has none”.
• Step-by-step walkthrough opens the key guide on reverbot.ai.
• Under Recommended, tap Gemini to use it, or tap the sliders button (three short lines with knobs) beside it to set it up.
• Under Advanced, tap Other AI providers to show Claude, ChatGPT and Grok. Without a key, a provider reads “Needs a key before it can answer”, and tapping it opens its setup.
• Answer speed, AI for each feature and Video AI — Twelve Labs are further down.

The AI in use is tinted and ticked.

Get a free Gemini key

One Gemini key answers questions, speaks with Gemini voices and makes pictures and videos in Image & Video AI.

1. Open Settings › AI & API Keys.
2. Tap the sliders button beside Gemini.
3. Under How to get one, tap Get your Gemini key. Google AI Studio (aistudio.google.com/apikey) opens in Safari; sign in with any Google account.
4. If you’re new to AI Studio, Google has already made a key on the API Keys page. Otherwise, tap Create API key and choose a project, or Create project.
5. Tap the copy icon next to the key.
6. Return to Reverbot and paste the key into Gemini (Google) API Key.

Back on AI Provider & Keys, the summary at the top shows Gemini answering “On your own key”. After your trial starts, Get your key — required later above the big button opens a guided version of these steps.

Note: On Google’s free tier, Google may use what you send to improve its products, and people at Google may read it, except in the EEA, Switzerland and the UK. Turning on billing for the key’s project puts it on Google’s paid terms.

Add ChatGPT, Claude or Grok

These answer only on a key of your own, and each company bills you for every answer.

1. Open Settings › AI & API Keys.
2. Under Advanced, tap Other AI providers.
3. Tap the sliders button beside ChatGPT, Claude or Grok.
4. Under How to get one, tap the link, then make and copy a key on the company’s page.
5. Paste it into the key field, for example Claude (Anthropic) API Key.
6. Tap Use [provider] for answers, for example Use Claude for answers.

• OpenAI: Get your ChatGPT key opens platform.openai.com/api-keys. Add a payment method or credit under Settings › Billing, then tap Create new secret key under API keys.
• Claude: Get your Claude key opens platform.claude.com/settings/keys. Buy credits under Billing, then tap Create Key under API keys.
• xAI: Get your Grok key opens console.x.ai. Add credits under Billing, then tap Create API Key under API Keys. The key starts with “xai-”.

Change a provider’s settings

Open Settings › AI & API Keys and tap the sliders button beside a provider. Its screen, for example Claude (Anthropic), shows:

• What it gives for free, if anything.
• Use [provider] for answers, which makes it your default AI. It reads “Selected as your default AI” when it is.
• The key field, with Show the key and Remove this key.
• Model: the AI model it answers with. The quickest is marked “Recommended”. Models marked “· from your key” are other models your key can use, and aren’t speed-rated.
• How to get one: the steps, and the link to the page where the key is made.

If you leave a provider without a key, Reverbot goes back to Gemini when you close its screen. A feature with its own AI keeps using it.

Choose an AI for each feature

Each feature can use its own AI, for example Claude for Questions and Gemini for translation, while the rest keep the default.

1. Open Settings › AI & API Keys.
2. Under AI for each feature, tap a feature, such as Questions. The AI for Questions screen opens.
3. Tap a provider, or tap the Default row to use the AI chosen at the top of AI Provider & Keys.
4. Optional: Under Model, choose a model. Recommended uses that provider’s quickest model.

A provider marked “— no key” needs its key first. Until you add it, the screen says, for example, “No Claude key is saved, so this feature uses the default until one is added.”

Tap Set up [provider] to add the key. A mode’s settings show the same choice as AI.

• Translation offers only Reverbot (included), Gemini and ChatGPT. On your key, each sentence, including other people’s voices, is one request billed to you by Google or OpenAI.
• Custom Conversation scenarios choose their AI in the scenario editor, and souls on their AI & voice page.

Note: On a phone call with Both of you, Reverbot always uses its own AI and translation.

Change answer speed

In Settings › AI & API Keys, under Answer speed, choose how much the AI thinks before it answers:

• Fastest, the default: the least thinking the AI allows. Best for spoken answers.
• Thoughtful: the AI thinks first. Slower to start, and can be better on hard questions.

When ChatGPT or a supported Claude model answers anything, OpenAI fast mode or Claude fast mode may appear below. Each asks for faster processing, which can cost more on your account; Claude fast mode also needs access on your Anthropic account.

Note: Answer speed doesn’t change how long Reverbot waits after you stop talking, which is Silence duration before answering, or how fast the voice talks.

Add a key for voices

Reverbot Voice and On-device (Apple) need no key. For another voice on your own account, open Settings › Voice & Speed and, under Voice Engine, tap one:

• Your own Gemini key, Your own OpenAI key or Your own xAI (Grok) key: the same keys as answers. Paste one in the field shown if it isn’t saved yet; the links there read Get your Gemini key, Get your OpenAI key and Get your xAI key.
• Your own ElevenLabs key: paste it in ElevenLabs API key. Get your ElevenLabs key opens elevenlabs.io/app/developers/api-keys; tap Create API key. A free account works. If you limit the key, allow Text to Speech, Speech to Speech and Voices.
• Your own Google key: Google Cloud voices. Paste it in Google Cloud API key under Google Cloud TTS, then choose a Quality tier. How to get a key walks you through Google Cloud (console.cloud.google.com): create a project, link billing, tap Enable on the Text-to-Speech page, then make a key restricted to the Cloud Text-to-Speech API. This key is only for voices.

Tap Test voice after pasting. Each company bills you past any free allowance it gives.

Add a key for faces

A personal face talks for free on your iPhone. For a realistic face on your own account, add an Anam or Simli key. Record a personal face first, as the big button topic explains.

1. Open Settings › Center button appearance.
2. Under Your default style, tap Personal face.
3. Tap the recorded face, then tap How [name]’s face talks.
4. Paste the key in Your Anam API key or Your Simli API key.
5. Tap Save Anam key or Save Simli key.

• Anam, for a live face: Get your Anam key opens lab.anam.ai/api-keys. Sign in, tap Create API key and copy it. Anam charges your account for each minute a live face runs.
• Simli, for a clip of each sentence: Get your Simli key opens app.simli.com/apikey. Sign in and copy the key shown. Simli charges for making a face and for each animated sentence.

The screen shows the cost and terms before anything is uploaded.

Add a key for Voice Changer

Voice Changer’s AI voices convert a recording on your own account.

1. Open Settings › Voice Changer.
2. Under Natural AI voices · your own account, choose ElevenLabs, Respeecher or Resemble.
3. Paste the key in the field below. For Respeecher or Resemble, tap Save key.

• ElevenLabs uses the same key as voices.
• Respeecher: Get your Respeecher key opens marketplace.respeecher.com/account. Sign in to Respeecher Marketplace and generate an API key with an expiry date. A key from Respeecher’s text-to-speech playground won’t work.
• Resemble: Get your Resemble key opens app.resemble.ai. Sign in, open Account › API and copy your API token. Resemble also needs your own Cloudflare storage.

Each conversion is charged to your own account.

Add a picture or video key

BytePlus makes Seedream pictures and Seedance videos:

1. Open Settings › Image & Video AI.
2. Under Advanced account setup, tap Get your BytePlus key. BytePlus (console.byteplus.com) opens; sign up and verify your identity.
3. In ModelArk, turn on Seedream for pictures and Seedance for videos.
4. Under API Keys, tap Create API Key and copy the key.
5. Back in Reverbot, paste the key in BytePlus API key.
6. Tap Save BytePlus key.
7. Choose the BytePlus key region the key was made in. Seedance needs an Asia Pacific key.

The same screen takes your Gemini, xAI and OpenAI keys, with the links Get your Gemini key, Get your xAI key and Get your OpenAI key.

Twelve Labs lets souls watch whole videos. In Settings › AI & API Keys, under Video AI — Twelve Labs, open How to get a key and tap Twelve Labs: API keys (playground.twelvelabs.io). Sign up free, tap Create API Key, then paste the key back in Reverbot and tap Save the key.

Each picture, video or watched video is billed to your account, and its cost shows before anything is made.

Set up Cloudflare

Video calls and recordings in Chat, Resemble conversions and souls’ photos and videos online run on your own Cloudflare account. Its guided setup makes the key for you.

1. Tap the Chat button in the top bar, then tap Settings.
2. Under Calls & meetings, tap Video calls & recordings.
3. If you have no Cloudflare account, tap Sign up at Cloudflare.
4. Tap Create my Reverbot key. Cloudflare opens in Safari with everything filled in.
5. On Cloudflare’s page, tap Continue to summary, then Create Token.
6. Copy the key. Cloudflare shows it only once.
7. Return to Reverbot and tap Paste.
8. Tap Connect.

Cloudflare bills your own account past its free allowance; Reverbot never charges for it.

If a key doesn’t work

• If a message says a company “rejected your API key”, copy the whole key again with the company’s copy button and paste it once more, or make a new key.
• If it says the account “has no credit left”, add credit or a payment method in that company’s billing settings.
• If you see “Google rejected this Gemini key.”, make a new key in AI Studio. Google refuses older keys that aren’t restricted to the Gemini API; a new key from AI Studio already is.
• If a company “is rate-limiting you”, wait: per-minute limits clear in seconds, and a daily one resets the next day.
• If a company “doesn’t recognise the selected model”, choose another Model on its screen.
• If a pasted key doesn’t seem to take effect, go back from its screen: a new key counts once you leave it.

Revoke keys you no longer use on the company’s page. Still not working? In Settings, under Feedback, tap Contact Support, and never include the key itself.

Sign in to the same account

1. On your iPhone, open Settings. Under Profile, the line under your name shows how you signed in, for example “Signed in with Apple”.
2. On reverbot.ai, open Manage App.
3. Sign in the same way: Continue with Apple, Continue with Google, or your email and password, then Log in.
4. Open Reverbot on your iPhone so its settings and workspaces sync.

Note: Signing in another way can open a different account. A guest account can’t sign in on the website: in the app, open Settings and, under Account, tap Create an account or sign in first.

Change settings from your computer

1. In Manage App, find the setting under Settings.
2. Change it, then click Save changes.
3. Wait for “Saved — your phone will pick this up next time you open the app.”
4. Open Reverbot on your iPhone. The change arrives when the app comes to the foreground.

Changes you don’t save never reach your iPhone.

Upload a reference document

1. In the app, set up the Job or Personal Assistant workspace you want the document in, and let it sync.
2. In Manage App, under Documents, choose that workspace in Workspace.
3. Under File, choose a PDF, TXT, MD or RTF file smaller than 20 MB.
4. Click Upload. The file appears under Your uploads as “Waiting for your phone”.
5. Open Reverbot on your iPhone. It imports the document and reads its text, and the upload then shows “On your phone”.

Check the document in its workspace before you ask questions about it.

Remove a document

1. In Manage App, find the document under Your uploads.
2. Click Remove upload, then confirm.

Note: This removes the website’s copy only. A copy your iPhone already imported stays there: delete it in its workspace in the app.

Keys and signing out

Add AI keys on your iPhone, in Settings › AI & API Keys. Manage App never shows them.

When you finish on a shared computer, click Sign out. This signs out the browser, not your iPhone.

Delete the account or stop billing

Important: Deleting your account can’t be undone. Your settings, uploaded documents, question history and phone-call balance are erased.

1. Save anything you want to keep.
2. On your iPhone, open Settings, scroll to the end and tap Delete account, then confirm.
3. To stop a subscription renewing, cancel it with Apple: in Settings, under Subscription, tap Manage subscription.

Deleting the account or the app doesn’t cancel an Apple subscription.

If an upload doesn’t arrive

• If Documents says “Open the app once to sync your workspaces, then upload here.”, open Reverbot on your iPhone while signed in, then reload Manage App.
• If an upload shows “Waiting for your phone”, open Reverbot on your iPhone with an internet connection.
• If it shows “Workspace not in your current list”, that workspace is no longer in the app: choose another one and upload the file again.
• If uploading stops with “Upload failed”, check your connection and try again.
• If it shows “Failed”, read the message under it, check that the file is a PDF, TXT, MD or RTF file under 20 MB, and upload it again. A scanned image with no text gives Reverbot nothing to read.