Your recordings,
in words.
Turn a video into a transcript you can actually use.
Choose Qwen or Whisper, running on your machine.
Start with a recording
One link. A little processing. Every word.
Bring your recording
Paste a shared link or upload a file.
Let the local model listen
Your machine does the transcription.
Take your words with you
Copy or download the transcript.
Time labels mark audio section boundaries (up to 3 minutes), not word-level alignment. Review AI-generated text for accuracy.
Your local setup.
A few connections. All transcription stays local.
Speech recognition engines
The first transcription downloads model weights. Following runs reuse them. Configure a local model path for offline use.
Google Drive
Sign in with the Google account the video was shared with. The owner must allow downloading.
Connect Google Drive ↗YouTube & Zoom access
For recordings requiring sign-in, set YTDLP_COOKIES_FILE in your local .env to a Netscape-format cookies file. Enter Zoom recording passcodes under Recording access.
When a provider prevents downloads, download the recording through its interface and use Upload a file.