Skip to content
GigAI Tools
3 tools
Problem → solution

Transcribe audio without sending it to a server

An interview under embargo or a client call under NDA shouldn't visit a transcription company's servers.

The quick answer

GigAI Video Intelligence runs OpenAI's Whisper model inside your browser via transformers.js, so the recording is transcribed on your own device and never uploaded. Otter, Rev and Fireflies all require upload; this doesn't. Free, no account - after the one-time model download it even works offline.

How to fix it

Follow these steps and you'll have it sorted in a couple of minutes.

  1. 1

    Drop the recording

    Audio or video - decoding and transcription both happen locally.

  2. 2

    Pick a model size

    Fast English (42 MB) for quick drafts; the multilingual model for accents and other languages. Downloads once, then caches.

  3. 3

    Edit and export

    Fix names inline, label speakers, then export TXT, Markdown, SRT or VTT.

Common mistakes to avoid
  • Assuming 'private' cloud transcription means unprocessed - upload IS processing on someone else's machine; read their retention terms.
  • Using the large model for a clear English recording - the small model is dramatically faster and usually enough.

Frequently asked questions

Ready? Start with Video Intelligence

Free, private and instant: everything runs in your browser.