Skip to content
unottr
Buy unottr

For researchers

Interviews that stay on your machine and stay findable.

Interview recordings often carry consent terms. unottr transcribes locally and lets you trace every claim back to the second it was spoken.

The workflow

  1. 01

    Transcribe a folder of interviews

    Drop recordings into a watched folder. Whisper runs locally, with a GPU when one is available.

  2. 02

    Separate speakers

    Automatic speaker identification labels who said what. Merge two speakers or reassign a segment when the split is wrong.

  3. 03

    Search every interview

    Full-text search covers every transcript. A hit opens the recording at that timestamp.

  4. 04

    Correct and cite

    Fix a transcript turn by hand, teach unottr recurring terms once, and use Ask to find answers with citations you can check.

AI features are optional. With a local model server such as Ollama, requests stay on your machine.

See pricingHow it worksJoin the beta

One-time purchase · $6.90

Your meetings stay yours.

Local transcripts, speaker labels and search, with optional AI that cites its sources.

See pricingJoin the free beta

Linux x64 · Apple Silicon · unottr 0.2.0 · MIT licensed