Speech to Text Online Free — Live Dictation in Your Browser
Runs 100% in your browser — your files never leave your device.
Tap the microphone, talk naturally, and watch your words appear as text — a live dictation pad powered by a speech model running entirely in your browser. Record in segments to build one transcript, then copy or download it. Nothing is uploaded: the model runs on your device after a one-time download, so dictation works offline. Ideal for drafts, notes, and anywhere typing is slower than talking.
How it works
- 1
Tap the microphone
Allow microphone access and start speaking — the timer shows how long you have been talking.
- 2
Tap stop
The on-device model transcribes what you said and appends it to the transcript. The first run includes a one-time setup.
- 3
Keep dictating or save
Record again to add more to the same transcript, then copy or download the finished text.
About this tool
How live dictation works here
Your microphone signal is recorded by the browser itself, decoded, and transcribed by the Whisper speech model running entirely on your device — the same engine the audio-to-text tool uses. Each recording segment is transcribed on its own and appended to the running transcript, so you can pace yourself: dictate a paragraph, check it, record the next. The model downloads once (about 45 MB) and is cached, after which dictation works even offline.
Dictating effectively
Speak punctuation where it matters — "period", "new paragraph", "comma" — and the transcript reads far cleaner, because the model transcribes words as it hears them. Short segments beat long ones: a paragraph at a time gives you a natural checkpoint to catch mis-hearings while the context is fresh. Background noise and cross-talk are the main accuracy enemies; a consistent distance from the microphone matters more than an expensive microphone.
Where dictation beats typing
Most people speak at 120 to 160 words per minute and type at about 40 — a first draft by voice is roughly three times faster, and the gap widens for longer text. Dictation shines for first drafts, emails, notes, and outlines: getting the words down is the hard part, and editing existing text is easier than producing it. For precision work like code or heavy formatting, typing still wins — dictation is for prose.
Privacy and limits
Everything happens on your device: the recording is held in browser memory, transcribed locally, and never uploaded. The transcript lives only in the page until you copy or download it — closing the tab clears it. The model is Whisper tiny, tuned for quick loading: clear speech in a quiet room transcribes very well, while heavy accents, crosstalk, or dense technical vocabulary may produce mis-hearings you fix by re-recording the segment.
Frequently asked questions
- Is my voice uploaded anywhere?
- No. The speech model runs in your browser after a one-time download. Your recording is transcribed on your own device and never sent anywhere.
- Can I dictate in multiple sessions?
- Yes — every recording appends to the same transcript until you clear it. Dictate a paragraph, review it, then record the next.
- How do I add punctuation?
- Say it: "comma", "period", "new paragraph". The model transcribes words as it hears them, so spoken punctuation lands in the text.
- Does dictation work offline?
- After the one-time model setup completes, yes — recording and transcription both run locally without an internet connection.
- Why did a word come out wrong?
- Background noise, distance from the microphone, and crosstalk are the usual causes. Re-record just that segment — each recording is transcribed on its own, so a redo does not disturb the rest of the transcript.
- Which browsers support it?
- Current Chrome, Edge, Firefox, and Safari all support the microphone capture and on-device model. A browser refresh clears the transcript, so copy long work before reloading.