How to Transcribe Speech in Real Time, Free
Updated 2026-06-21
To transcribe speech in real time, open a tool that runs a speech-to-text model on your own device and feeds it your microphone as you talk. With Live Transcription (beta), you grant mic access, start speaking, and words appear on screen within a second or two — no upload, no account, nothing sent to a server.
Start a live transcript in three steps
The flow is built for one thing: getting words on screen fast.
- Open Live Transcription (beta) and allow microphone access when the browser prompts you. The audio is processed locally, so it never leaves your machine.
- Wait a moment on the first run while the on-device model loads. This is a one-time cost per session — after it's ready, transcription keeps up in near real time.
- Start talking. Text streams in as you speak, so you can read the transcript building line by line and catch mistakes immediately.
Because this is a beta, treat the output as a fast draft rather than a flawless record. It's excellent for following along live and capturing the gist; for a verbatim, archival transcript you'll want to read through and fix the occasional missed word.
Get a more accurate live transcript
Real-time speech-to-text is sensitive to how you speak and what your mic hears. A few habits noticeably improve accuracy:
- Speak at a steady, natural pace. Rushing or trailing off mid-sentence is the most common cause of garbled output.
- Get close to a decent mic. A headset or laptop mic in a quiet room beats a phone across the table. Background chatter and echo are what trip the model up most.
- Reduce competing noise. Close a noisy window, mute notifications, and avoid talking over music or a TV.
- Enunciate names and numbers. Proper nouns and figures have no surrounding context to lean on, so they're the first things to come out wrong — say them clearly.
If a phrase lands wrong, keep going; you can correct the text afterward rather than stopping the flow.
When real-time beats recording first
Live transcription shines whenever you need the words while they're being said, not an hour later:
- Captions for accessibility — read along in real time during a call, a video, or a conversation when audio alone is hard to follow.
- Meetings and lectures — watch the transcript build so you can glance up for what you missed instead of frantically typing notes.
- Dictation and brain-dumps — talk out a draft, an email, or a stray idea and let the text appear as you think.
- Interviews — keep a live running record you can scan on the spot.
Why on-device transcription matters
Everything here runs in your browser using an on-device speech model, so your voice — and whatever you're discussing — stays on your computer. There's no upload, no signup, and nothing stored on a server to delete later. That makes it safe for sensitive calls, private notes, and confidential meetings where uploading audio to a cloud service isn't an option. Close the tab and it's gone.
Ready to see your words on screen as you speak? Open Live Transcription (beta), allow your mic, and start talking.