Speech to Text

How to use

1. 1Open this page in Chrome, Edge, Brave, Opera, or Safari. Firefox does not ship the Web Speech API and is shown a clear unsupported message.
2. 2Pick the language that matches your accent (for example en-US, en-GB, es-MX, fr-FR, hi-IN, zh-CN). Toggle continuous recording, interim results, and auto-punctuation on the right.
3. 3Click Start recording and allow microphone access when your browser prompts. Speak normally; the transcript fills in as the engine finalizes phrases.
4. 4Edit the transcript directly in the text area to fix any word the engine got wrong. Use New paragraph between sections and Auto-punctuate now to re-run capitalization and end punctuation.
5. 5Click Stop recording when you are done. Use Copy transcript or Download .txt to save the result, or Clear transcript to start a fresh dictation session.

About this tool

Speech to Text turns your voice into editable text using the Web Speech API that ships with modern browsers. Press Start recording, allow microphone access once, and your words stream into the transcript area as you speak; pause briefly and the engine commits a final phrase, which is then appended to the running text. Continuous mode keeps the recognizer alive across long pauses by automatically restarting on the silence end event, so you can dictate a long article without clicking Start again every minute. Interim results show partial guesses in real time so you can see how the engine is hearing you, and the live transcript area is fully editable, which means you can fix a misheard word with the keyboard without stopping recording. A language picker covers 40+ BCP47 locales including US, UK, Australian, Canadian, and Indian English, Latin American and European Spanish, French, German, Italian, Brazilian and European Portuguese, Dutch, the Nordic languages, Polish, Czech, Slovak, Hungarian, Romanian, Turkish, Ukrainian, Russian, Greek, Arabic, Hebrew, Hindi, Bengali, Tamil, Telugu, Thai, Vietnamese, Indonesian, Malay, Simplified and Traditional Chinese, Cantonese, Japanese, and Korean; picking the most specific tag for your accent often improves accuracy because the browser uses a matching acoustic model. The auto-punctuate option capitalizes sentence starts and adds a trailing period when one is missing so the transcript reads cleanly, and dedicated buttons re-run punctuation on demand and insert a new paragraph when you start a fresh section. When you are done, copy the transcript with one click or download it as a .txt file; the running word, character, and estimated reading time counts update live. Clear errors guide common issues: a no-microphone-found message when no input device is available, a not-allowed message that explains how to grant permission in site settings, and a network message when the speech service is unreachable. Browser support covers Chrome, Edge, Brave, Opera, Samsung Internet, and Safari (desktop plus iOS 14.5+); Firefox lacks the SpeechRecognition API and gets a clear unsupported notice rather than a broken experience. Useful for taking lecture notes hands-free, drafting blog posts and emails by voice, writing code comments while reading documentation, capturing meeting recaps and voice memos as searchable text, accessibility for users with limited mobility or typing fatigue, language learning practice with immediate visual feedback on pronunciation, and any task where talking is faster than typing. The audio and transcript do not pass through this site, but in Chromium-based browsers the Web Speech API forwards audio to Google's speech service for recognition and Safari uses Apple's own pipeline, which the tool surfaces in the privacy notes so the trade-off is clear.

Free to use. Works in your browser. No signup, no login.