OGG and Opus voice notes to text
OGG and Opus are what messaging apps use for voice notes. Drop one in and read it instead of listening to it — without forwarding a private message to a transcription service.
Opus is built for speech at low bitrates, so these small files carry more usable detail than their size suggests.
Language and translation need the Multilingual model.
Drop an audio or video file
or , paste with Ctrl + V, or
- MP3
- MP4
- WAV
- M4A
- MOV
- WEBM
- OGG
- FLAC
- AAC
The file is read by this page and never uploaded — no account, no size limit but your own RAM.
Preset: balanced English model. Switch to Multilingual for other languages.
Download
Click a line to jump the audio there. Click the text to fix a word — your edits go into every download.
How it works
- Open this page — the tool is already set up for “OGG to text”.
- Drop the file in, browse for it, or paste it with Ctrl+V.
- The model runs on your own device; the transcript appears as it goes.
- Correct anything misheard, then download text, subtitles or notes.
Voice notes are somebody else’s words
A voice note was sent to you, not published. Uploading it to a web service to get a transcript quietly involves a third party in a private conversation — one the person who recorded it never agreed to. That is the case regardless of whether the service is well-behaved, because the disclosure has already happened by the time the file arrives.
Doing it in your own browser removes the question. The file is read locally, the model runs locally, and nothing is transmitted. For a long voice note from a colleague, a family member or a client, that is the difference between reading a message and republishing it.
Getting the file out of a messaging app
On WhatsApp Web or Desktop you can download a voice message to disk; on mobile, share it to Files. Telegram allows saving voice messages directly. Both produce .ogg or .oga with Opus inside, which browsers decode natively. Signal, deliberately, makes exporting harder — which is consistent with its design.
Some apps hand you a file with no extension. Drop it in anyway; the container is detected from its contents rather than its name.
Other languages
Voice notes are often the most multilingual audio anyone deals with. Switch to the Multilingual model and either pick the language or leave detection on — it identifies the language from the first few seconds. If you want the sense rather than the exact words, set the output to translate and you get English text from speech in any of the supported languages.
Quality varies considerably by language. Spanish, French, German, Italian and Portuguese are strong at this model size; lower-resource languages are noticeably weaker, and short clips give language detection less to work with.
Things that save a re-run
- Very short notes give language detection little to go on — pick the language explicitly if it guesses wrong.
- Opus files are small; there is no need to convert them before transcribing.
- For a long rambling note, use the timestamped export so you can jump to the part that matters.
Read the transcript against the audio before you rely on it. Every speech model — this one and the paid cloud ones — mishears names, numbers and crosstalk, and it does so confidently. The player and editor here exist so that check takes minutes rather than an afternoon.
Frequently asked questions
Can I transcribe a WhatsApp voice message?
Yes, once you have saved it as a file. WhatsApp voice notes are Opus in an OGG container, which browsers decode natively. Download it from WhatsApp Web or share it to Files on mobile, then drop it in.
Is the voice note sent anywhere?
No. It is read by the page and processed in your browser. Nothing is uploaded, which is the point — a voice note is a private message, not something to hand to a third-party service.
What if the note is not in English?
Switch to the Multilingual model. It detects the language automatically, or you can pick one, and it can also translate the speech straight into English text.
The file has no extension — will it still work?
Yes. The browser identifies the format from the file contents rather than the name, so a mislabelled or extension-less file usually decodes without any trouble.