Someone speaks, and a second later their sentence is on screen, transcribed, cleaned of the ums and false starts, and translated into every language you picked, all at once. When the call ends, the whole conversation is written down, ready to keep.
New accounts are free: 5,000 words a week, no card.
Nothing to install, and nothing for the other people on the call to do. It runs in the browser on the laptop or phone that is already in the room. Do tell the others it is running: the translation is there for everyone to read, and some places require that everyone in the conversation knows.
Choose what to translate into, one language or five, and what kind of call this is: a meeting, a family catch-up, a book club. The spoken language is detected on every sentence, so nobody has to say who is talking.
A speech detector in the browser notices when someone starts and stops, so nothing is sent while the room is quiet and nobody holds a button. A long monologue is split at its natural pauses.
Each sentence appears in every chosen language within a second or two, and can be read aloud. When the call ends, export it as a subtitle file.
A meeting you recorded, a voice memo, a video of a talk: upload the audio or video file and it is transcribed and translated into every language you pick, sentence by sentence, as if it were being spoken into the microphone. When it is done, download the whole transcript with a timestamp on every sentence.
Tap the language badges on the menu bar to open Settings. Pick what to translate into, one language or several, and the kind of call. The spoken language is detected for you.
It is on the menu bar, next to Settings. Choose an audio or video file: MP3, WAV, M4A, Ogg or MP4. From a video, only the sound is used.
Each sentence appears with its position in the recording, the original first and every translation under it. When it finishes, Export in Settings downloads the timestamped transcript as a subtitle file.
Uploading needs a free account. The words in a recording count the same as words spoken live.
Four things, in order, to every sentence anyone says.
Whoever is talking, in whichever of 25 languages, is written down as they speak. The language is detected sentence by sentence, so a call that switches between English and Mandarin needs no settings changed mid-way.
The ums, the false starts, the sentence begun twice: all gone from the transcript. When someone corrects themselves, you get what they settled on. Numbers, names and dates are never touched in the cleaning.
One sentence, every language you picked, all streaming at once. Pick the kind of call, whether a meeting, a formal event, friends, politics, a book club or engineering, and the register follows: formal stays formal, casual stays casual, in each language’s own way.
Every sentence is kept with the time it was said, in the original and in every language. It survives a reload, and exports as a subtitle file for the recording or a text file anyone can open.
Decisions get made in the last five minutes, when half the room has stopped taking notes. Every sentence is kept as it is spoken, in the original and in every language it was translated into, with the time beside it.
Each entry keeps what was said and how it was translated, so a number or a date can be checked against the original rather than against memory.
The transcript survives a reload and stays until it is cleared. Export writes it as an SRT file: subtitles for the recording, or plain text anyone can read.
A colleague in another time zone, a relative who missed the call: they can read exactly what was said, in their own language.
“We needed something for our face-to-face meetings, not just video calls. Meeting Translator is easy to use and translates the conversation live, right there in the room.”
Both plans are the same translator, with every language, every call type and every rule. The difference is how much can be said.
$0 / forever
No card and no trial clock. Signing up puts you here.
Everyone who speaks draws on the same weekly words.
$20 / month
Billed monthly through Stripe. Cancel any time.
You can upgrade from inside the app once you have an account.
The words actually spoken, whoever said them. Translations are free: a sentence costs the same whether it goes into one language or five, and hearing it read aloud costs nothing. Chinese, Japanese and other languages written without spaces are counted by character.
People speak around 130 to 150 words a minute, so 5,000 words is over half an hour of actual talking: a full meeting, or several short calls. If you run out mid-call the app says so, and the allowance comes back over the following week.
Usage is counted from right now backwards over seven days rather than reset on a fixed day. Words spent in Monday’s call stop counting the next Monday, so the allowance returns a little at a time. There is no reset hour to wait for.
No. One person runs it, on the device that hears the room or the call audio; everyone else just talks. Share the screen into the call and the others see the translations too.
Speech goes to the third-party speech recognition and translation services this deployment is set up to use, and the text comes back. This server keeps neither the audio nor the transcript: the call stays in your browser until you clear it or export it. Your account holds your email address, whether you subscribe, and how many words each sentence used. An uploaded recording is the one exception: the file is held on this server while it is being transcribed, and deleted when that finishes or shortly after.
With your email address. We send a link, you tap it, and you are in. There is no password to choose or forget, the same link creates your account the first time, and each link works once.
Yes. Subscriptions are handled in Stripe’s billing portal, which you reach from inside the app, and cancelling takes effect without asking anyone. Your account goes back to the free plan rather than disappearing.
An SRT subtitle file: every sentence with the time it was said, the cleaned-up original on one line and each translation on its own line beneath it. Drop it onto the call recording as subtitles, or open it as text. Nothing is uploaded to this server for that to work.
Through the device’s microphone, the same way it hears a room. On a video call, let the other side come out of the speakers rather than headphones, and put the laptop where it can hear both them and you. In a meeting room, one laptop or phone on the table is enough.
Yes. Tap the upload icon on the menu bar, next to Settings, and choose an audio or video file: MP3, WAV, M4A, Ogg or MP4. It is transcribed and translated sentence by sentence, the same as a live call, and Export gives you the subtitle file with a timestamp on every sentence. Uploading needs an account.
An email address is all it takes, and you will be set up in under a minute.
Start free