Speech to Text
How to Use
🎙️ Speech to Text
This tool converts spoken words into text. Press start, talk into your microphone, and the transcript appears in the text box as you speak; press stop when you are done. Recognition relies on the speech engine built into the browser, supports Chinese and English, and lets you pick the language from a dropdown.
The transcript stays editable, which matters because no speech engine is perfectly accurate. Proper nouns, personal names, and homophones regularly need a manual fix. Once you are happy with the wording, copy it with one click or download it as a txt file. The tool also keeps a history for the current session so you can scroll back through earlier segments.
It works well for cleaning up meeting notes, capturing an idea before it slips away, turning a spoken draft into written text, or dictating a long passage when typing is inconvenient. There is no hard limit on recording length, although very long single takes tend to slow recognition down.
Audio is processed by the browser speech service rather than by this page, so the microphone stream leaves your device only to that service and never reaches our server. If your browser has no speech engine, or the service cannot be reached, the tool tells you instead of producing a fake transcript.
Keep sentences short and paced normally: dictating in long unbroken paragraphs gives the engine fewer places to settle on a boundary, and the transcript becomes harder to read back. If a word keeps coming out wrong, spelling it out or slowing down usually works better than repeating it.
✨ Features
- Uses the browser native speech engine, with Chinese and English available
- Transcript appears live, so text shows up while you are still talking
- The result is editable, and one click copies it out
- Download the transcript as a txt file for archiving
- Keeps a history for the current session so earlier segments can be reviewed
- Clear recording state indicator, so you always know whether it is listening
🚀 How to Use
- Press the start button; the browser asks for microphone permission, so choose allow
- Pick the recognition language from the dropdown (Chinese or English)
- Speak normally into the microphone and watch the text fill in
- Press stop when finished, then read through and correct any mistakes
- Copy the text, or download it as a txt file
⚠️ Notes
- Depends on the browser speech service: recognition is performed by that service and requires network access; Chrome and Edge work well, while Safari and Firefox may not offer it at all
- Microphone permission is required: the first run triggers a permission prompt, and if you decline you must re-enable it in the browser site settings
- Accuracy follows your environment: a quiet room, clear pronunciation, and a closer microphone noticeably improve results
- Proper nouns need manual correction: names, places, and technical terms are often heard as homophones, so read the final text once
- Nothing is stored: history lives only in the current page and is cleared on refresh, so copy or download anything important