Aloud
Live captions, as it is said
Speech turns into text the moment it is spoken, split into turns so a back-and-forth reads like a conversation rather than one long paragraph.
LIVE live captions • 70+ languages • works offline
Everything said, in writing.
Put the phone on the table and Aloud writes out what the room is saying, as it is said, in text big enough to read from across it. Nobody else has to install anything, join anything, or hold anything. It is a hearing aid for a conversation you would otherwise have to keep asking people to repeat.
LanguagesAround 70
OfflineNearby needs no signal
AccountNone required
When it is not sure, it says so
Speech recognition guesses. Every other app hands you the guess as a finished sentence and lets you believe all of it equally — which is how a misheard name, dose or time becomes something you find out about later, after acting on it.
Ask for Doctor Eefa when you get there.
Apple’s recogniser returns a confidence for every word, and sometimes the other readings it weighed and threw away. That is normally discarded somewhere between the framework and the screen. Aloud keeps it, and compares each word with the rest of its own turn — so a word it hesitated over is drawn as hesitated over.
- 1 The word is marked, not slipped past youHeld back and underlined, so you can see which parts of the sentence to trust.
- 2 Tap it to see what else it heardThe readings it also considered. Pick the right one and the sentence is corrected — in the saved transcript and its summary, not just on screen.
- 3 Or turn the phone round and askJust that phrase, held up with “could you say this bit again?” — so they repeat three words rather than the lot.
- 4 The correction sticksA name you fix once is given to the recogniser as a hint from then on. The names that get misheard are the ones that come up again.
- 5 And it is still there afterwardsOpen a saved session days later and it says how many words were guesses, and still lets you answer them.
It does not catch everything, and it would be worth less if we pretended it did: the recogniser can be wrong and sure of itself, and no amount of reading its confidence catches that. What this removes is the far more common case — the word it was never confident about in the first place, handed to you looking exactly like the words it was.
Built for the conversation in front of you
Not a meeting recorder you read afterwards. Everything here is pointed at the words landing on screen fast enough to reply to.
A colour per speaker
In Studio, each change of speaker takes its own colour, so a three-way conversation stays readable instead of blurring into one voice.
Write back
Your half of the conversation. Type a reply at full screen size and hold the phone up, or drop it straight into the transcript mid-session.
Catch me up
Attention drifted? One tap summarises the last minute. It reads the transcript already on your phone — nothing is sent anywhere to do it.
A timer that offers you more
Set a session to stop after 5 to 90 minutes. When the clock runs down it pauses and asks — plus five, plus fifteen, or finish — rather than closing on you mid-sentence.
Every session saved
Sessions are filed automatically with a title and summary, searchable by any word anyone said, and stored on your phone alone.
Three ways to listen
Not marketing tiers — each one sets the speech recogniser up differently, and the right choice depends on how far away the voice is.
Nearby
FreeRuns the speech model on the phone itself. Best within arm's reach — on-device models start dropping words across a room.
- ✓Works with no signal at all
- ✓Nothing is sent anywhere
- ✓Free and unlimited, no account
Range
TrialHands the work to Apple's server recogniser, which hears markedly further and copes better with a noisy room.
- ✓Picks up voices across a room
- ✓Cleaner punctuation
- ✓Uses your own saved vocabulary
Studio
TrialEverything Range does, plus a shorter gap before a turn is called finished — so a change of speaker lands on its own line.
- ✓A colour per speaker
- ✓Splits turns on short pauses
- ✓Sixty-second catch-ups
Three steps, thirty seconds
There is no account to make and nothing for the other person to do.
- 1Put the TV on, or sit down with someone.
- 2Hit Listen. Words start appearing straight away.
- 3Pinch the text to whatever size you can read comfortably.
Your conversations stay on your phone
There is no account and no server of ours. Saved sessions live in the app's own storage on your device, and Nearby never sends audio anywhere at all.
Read Privacy PolicyWho it is for
Anyone who has spent a conversation nodding along and hoping they guessed right.
Deaf and hard of hearing
Captions for the room, not just for a video. The text scales to whatever size you need, and Write Back gives you the other half of the conversation.
Noisy rooms and quiet voices
Cafés, hospital corridors, a relative who has grown softly spoken. Range hears across the table when your ears cannot.
Meetings and appointments
A session saves itself with a title and a summary, so what the doctor or the client actually said is there to check afterwards.