How to Transcribe Audio Online
Turn a recording into text you can read, edit, and reuse. You do not need to install an app or create an account.
On this page
1. Choose your recording
Use the clearest original recording you have. Common formats include MP3, WAV, M4A, and WebM. The queued upload limit is 1 GiB per file. Keep the page open until the upload finishes; uploaded jobs continue on the server.
For a voice message, save or share the audio file from your messaging app first. TalkToTextly only receives the recording you select; it does not open your chats or join meetings.
2. Start the transcription
Open TalkToTextly, select your audio, and check the language. Read the upload notice, then confirm the file. Audio is uploaded securely and sent through Hugging Face Inference Providers to DeepInfra for transcription. Audio is deleted after completion or error.
Processing time varies with recording length, queue demand and service availability. After upload, you can return in the same browser to recover the result for seven days. Keep your original file and download a copy of the transcript before clearing browser storage.
3. Review the words
Use the transcript as a first draft. Listen again around names, numbers, specialist terms, and unclear speech. Background noise and people talking over each other can cause mistakes.
Correct the text before quoting it or sharing it. For an important interview or document, compare the final wording with the recording.
4. Copy or export the result
Copy the words into your notes, download TXT for plain text or DOCX for a document, or choose SRT or VTT for subtitles. Review subtitle wording and timing before publishing.
Your transcript may remain in this browser's history. Delete it from the history when you no longer need it, especially on a shared computer.
Only upload recordings you have permission to share. Read the privacy information if your audio contains confidential material.
Transcribe your audio