Feature
Real-time transcription with live translation
See speech transcribed as it happens, with translated lines underneath and a transcript you can edit and export when the session ends.
TanLens transcribes speech as it happens and shows the translation under each finished line. The text settles into a transcript as the session runs, so you are not left with a disposable overlay when the call ends. Live sessions can use a microphone or a browser tab; uploaded files follow the same transcription pipeline after processing.
01What appears while people speak
The current HTTP path sends complete two-second audio chunks to the recognizer, then returns stable source text followed by the translation. The speech provider, translation provider and network all add to the real delay; TanLens has not published production p50 or p95 measurements yet. Long pauses, crosstalk and heavy background music make both recognition and translation less reliable, as they would for any live caption system.
02A transcript instead of a disposable overlay
A live transcript is useful only if you can fix it. After the session, open the transcript, correct names and numbers, and remove the false starts that make a recording hard to read. TanLens keeps a transcript only when you choose to save it; audio itself is processed for captions and is not stored by default. Saved transcripts can be exported as SRT, VTT, ASS or TXT, in bilingual or single-language form, which makes the same text usable for a video edit, a web player or meeting notes.
Frequently asked questions
Related pages
Put words on the screen tonight
Twenty free minutes a month, no card, no meeting bot. Start with your microphone, then add the extension for tab audio.