Speech to Text
Speech APILive speech-to-text transcription using the browser's built-in Web Speech API. No audio is recorded or stored by this page.
Audio Controls & Language
Microphone idle. Ready to transcribe.
Dictation Tips:
- Grant microphone permission when prompted by your browser.
- Speak clearly at a regular conversational pace.
- Say punctuation explicitly (for example: "period", "comma", "question mark").
Live Transcript
0 words · 0 chars
About This Tool
Wraps the browser's native SpeechRecognition API (including the webkitSpeechRecognition prefix) into a streamlined transcription workbench. Words appear live as you speak with interim guesses, solidifying into finalized text as phrases finish.
This runs as a browser capability rather than a server-side transcription service. Support and speech recognition coverage depend on your browser engine (Chromium browsers like Chrome and Edge are fully supported, while Firefox and Safari currently do not implement this API).
Key Features
- Upfront Support Detection: Immediate compatibility check prevents silent failures on unsupported browser engines.
- Multiple Languages: Dictate in English, Spanish, French, German, Japanese, Mandarin, Indonesian, and more.
- Live Interim Stream: Words render in real time as spoken audio is parsed, finalizing smoothly without stutter.
- Continuous Transcription: Maintains an active listening session across short pauses until you click Stop.
- Instant Export: Single-click copy to clipboard or direct .txt file download of your full dictation.
Frequently Asked Questions
Why does this say "not supported" in Firefox or Safari?
This tool uses the Web Speech API SpeechRecognition interface, which is currently implemented by Chromium-based browsers (Chrome, Edge, Brave, Opera) via the webkitSpeechRecognition prefix. Firefox and Safari do not implement it, so the tool will not work there: this is a browser limitation, not a bug in this page.
Is my audio sent to a server?
The Web Speech API in Chromium browsers sends audio to Google's speech recognition service for processing: that is how the browser API itself works, independent of this tool. No audio or transcript is sent anywhere by this page beyond what the browser's own API does internally.
What is the lighter or interim text in the transcript?
That is the interim result: the browser's current best guess for what you just said, which can still change as you keep talking. Once the browser finalizes a phrase it becomes permanent text and stops changing.
Can I change the recognition language mid-session?
Changing the language while actively listening restarts recognition with the new language, since SpeechRecognition only supports setting a language before it starts. Stop listening first if you want a clean switch.
Why did listening stop by itself?
Browsers typically end a recognition session after a prolonged period of silence or a maximum duration limit. Click "Start listening" again to resume: previously transcribed text in the editor is preserved.
