TypeTarget — Speech to text anywhere

Transcribe browser tab audio directly into any text field, with your choice of AI model for free.

TypeTarget listens to the audio playing in any browser tab — a video, a meeting, a lecture, a podcast — and types what is said, live, into any text field you choose, in any tab. Take notes from a lecture straight into your doc. Caption a meeting into a chat box. Pull a transcript of a video into a form, an email or a note-taking app. Or type into TypeTarget's own text box and save to your own device as a .txt file. HOW IT WORKS 1. Right-click anywhere on the tab you want to transcribe and choose TypeTarget → Listen to this tab. 2. Choose where the text goes, from the same right-click menu on any page: • Type to this field: right-click any text box (a doc, a chat, an email, a form). • Type to new file: a text box over the bottom of the page you are on. • Type to new tab: a full-page text box in a tab of its own. 3. Transcribed text appears there as people speak, after anything already in the box, without touching your own edits. Choose Stop typing or Stop listening from the same menu, or switch fields any time. The box you are typing into gets an outline with a listening timer and buttons to save its text as a .txt file or stop typing. A new-file box can also be minimized or moved into its own tab. The toolbar popup has a Start/Stop button for listening and one for typing, plus the model picker, the API key for each hosted model, the spoken language and live diagnostics. Turn off "Show outline" to type without any outline or buttons on the site; the toolbar icon shows REC instead. CHOOSE YOUR TRANSCRIPTION MODEL • Groq (Whisper Large v3 / v3 Turbo) — fast and accurate on any computer. Uses your own Groq API key. • Gemini 3.5 Transcribe Live (Google) — Google's streaming transcription. Uses your own Gemini API key. • Local Whisper, in your browser — Tiny, Base, Small, Medium and Large (v2, v3, v3 Turbo), 31 MB to 1 GB, downloaded once and cached. No key, no account, and audio never leaves your device. Keeps up live only on fast computers; TypeTarget tells you when it is falling behind. PRIVATE BY DESIGN • No accounts, no servers, no analytics, no tracking. • Audio goes only to the provider you pick, using your own key, and never to the developer. • API keys are stored on your device only. • Free and open source (MIT): https://github.com/wllmmh/typetarget LANGUAGES • Transcription: every model detects the language being spoken and types it in that language. The Whisper models, local and on Groq, cover Whisper's 99 languages. • Interface: TypeTarget's menus and buttons follow your browser's language, in Arabic, Bengali, Chinese (Simplified and Traditional), Czech, English, Filipino, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Marathi, Polish, Portuguese (Brazil and Portugal), Russian, Spanish, Telugu, Thai, Turkish, Ukrainian, Urdu and Vietnamese.