Speak into your microphone and watch it turn into text live, right on this page, using your browser's built-in speech recognition.
⚠ Heads up: unlike most tools on this site, this one is not fully local — see the FAQ below for how it uses your browser's speech recognition.
This tool is a thin, honest wrapper around the SpeechRecognition interface that ships inside modern browsers. It is not a custom machine learning model running on your device — it simply asks your browser to listen to your microphone and hand back text as it recognizes it.
Recognition is handled entirely by your browser's own SpeechRecognition engine, so there's nothing to download or install.
As you talk, a lighter "interim" guess updates in real time and locks in as finalized text once the browser is confident in it.
In most browsers, short audio snippets are sent to that browser vendor's cloud speech service to produce the transcript — unlike most tools on this site.
No — unlike most tools on My Panda Toolbox, this one is not fully local. It uses your browser's built-in Web Speech API for recognition. In most browsers, including Chrome and Edge, that API sends short snippets of your audio to the browser vendor's own cloud speech service to produce a transcript, then returns the text to this page. Nothing is stored by this site, but your audio does leave your device on its way to that cloud service. If you're dictating anything sensitive or private, keep that in mind.
Speech recognition works best in Chrome and Microsoft Edge on desktop. Firefox does not support the Web Speech API's recognition feature, and support on Safari and mobile browsers is inconsistent. If the Start button is disabled, your browser doesn't support it — try Chrome or Edge instead.
Pick a language from the dropdown before pressing Start listening. The language is applied when a new recognition session begins, so if you change it while the tool is actively listening, stop first, then start again for the new language to take effect.
This tool uses continuous mode, which keeps listening across natural pauses in your speech and keeps appending to the transcript until you press Stop. That's different from single-utterance mode, which some voice tools use, where recognition automatically stops after it detects the end of one sentence or phrase.
Browsers can end a speech recognition session on their own after a period of silence, a network hiccup, or when a microphone permission or vendor cloud call fails. This tool doesn't auto-restart for you, to avoid confusing restart loops — just press Start listening again to resume.
More handy tools from My Panda Toolbox.