About Timecode Transcript
Timecode Transcript turns recordings into transcripts with timecodes that people can actually use: brackets for notes and reports, SMPTE frames that line up with an edit timeline, and subtitles for video.
It is made and run by Zehong Liu, an independent developer. The speech recognition runs on Cloudflare Workers AI with OpenAI’s Whisper large-v3-turbo model. Cloudflare runs the model on its own servers; your audio is not sent to OpenAI. The decoding, the timecodes and the downloads are built here and run in your browser.
The tool is free for the first 10 minutes of up to 3 files a day. Longer files and the editing formats are paid for with minute packs: one payment, no subscription, and unused minutes refunded within 14 days. That is how the site pays for its servers.
Questions, bugs and ideas are welcome at the address on the contact page.