Skip to main content

Transcribe in your browser

You can transcribe files without an API key or any code: sign in, open the Transcribe page, drop in audio or video files and copy or download the text. It uses the same minutes and handles your audio the same way as the API.

Add files​

  • Drop files anywhere on the page, or press Choose files. You can add several at once; they are transcribed one after another.
  • Press Record to record from your microphone, then Stop and transcribe. Your browser asks for permission the first time.
  • Each file can be up to 100 MB and 3 hours long, in any of the formats the API takes. A larger file is refused before it is sent.

Choose what comes back​

  • Text, SRT subtitles or VTT subtitles (response formats). The choice applies to the files you add after changing it.
  • Speaker labels turns on speaker labels. Text then comes back as turns: Speaker A: ..., Speaker B: ....
  • The page names no language, so files are read in your account's default language: English, unless other languages are on for your account (see Languages on your account page).

Results​

Each file gets a card with Copy, Download and Remove. Downloads are named after the file, such as interview.srt.

Results stay in the browser tab until you close it. Turn on Keep results in this browser to keep the newest 20 after the tab closes, on this computer only. Clear results removes them. We don't store transcripts; as with the API, we keep one billing line per request.

How private it is​

Your audio is handled exactly as it is through the API: held in memory while it is transcribed, then erased. The difference is your browser. Extensions that can read web pages can also see the files you add and the transcripts on the page, and kept results sit in the browser's storage on your computer. The page says so the first time you open it. For the most private setup, use the API, or the PGP options below with files you encrypt yourself.

Expert mode: PGP encryption​

With PGP encryption on your plan, turn on Expert mode:

  • Encrypt audio in this browser before sending (on by default) encrypts each file to your account's own key before it leaves the page. Files that are already encrypted, such as a .gpg you made with gpg to your account's key, are sent as they are. The account key's fingerprint is shown under Expert mode.
  • Encrypt the transcript to one of your public keys returns the transcript encrypted to that key and signed by us, as an .asc download that you decrypt with gpg on your computer.
  • Paste a public key adds a key to your account, the same as adding it on the account page, including the notice email. Check that the fingerprint it shows matches gpg --fingerprint on your computer.

The page never asks for a private key and never decrypts anything. Encrypting in the browser protects the file on its way to us, but the page and any extension can still read the file before it is encrypted. For the strongest protection, encrypt the recording with gpg and add the .gpg file, with the transcript encrypted to your key.

Without PGP on your plan, Expert mode shows the plans that include it.

If something goes wrong​

The file's card shows the same message the API would send, for example when your minutes run out (errors). Try again sends the file again. If the page says it has been open too long, reload it. If you were signed out, sign in and add the file again.