For journalists and researchers

Interview transcription without the all-night typing session

A one-hour interview can take three hours to transcribe by hand, and most automatic tools still leave you fixing accents, names and overlapping speech. VoxNote is built for the part that actually costs you time: turning a long recording into quotes, notes and a transcript you can search.

Record directly in the browser or upload an existing audio file. Whisper Large v3 Turbo transcribes it with timestamps, then the AI layer produces a summary, action items and candidate key quotes you can check against the audio.

Start recording

Timestamps you can navigate

Every segment carries a start and end time, so you can jump back to the exact moment before publishing a quote.

Built for messy interview audio

Voice activity detection skips long silences, and the optional names prompt helps with people, places and technical terms.

Quotes, not just a wall of text

Ask for key quotes or a research summary instead of a generic summary that loses the details you care about.

Exports that fit your workflow

Download TXT, SRT, VTT or JSON. Paste the transcript into your notes app, or keep the structured segments for analysis.

How it works

1. Record or upload
Use your microphone for a new interview or upload an existing MP3, M4A or WAV file.
2. Transcribe
Choose the spoken language, add names or acronyms, and start the transcription.
3. Review the quotes
Use the summary and key quotes to find the moments worth publishing, then check them against the audio.
4. Export
Download the transcript or subtitle file, or save the session in your browser for later.

Frequently asked questions

Can it handle a two-hour interview?

Yes. Recordings are split into short segments in the browser and transcribed sequentially, so long interviews do not depend on one huge upload. Each AI request is capped at two minutes or five megabytes, and the free daily allowance still applies.

Does it support strong accents?

Whisper handles accents well when the audio is reasonably clear, but every automated transcript should be reviewed. Heavy background noise and overlapping speakers remain difficult for all ASR systems.

Can I use the transcript in an article?

The transcript is a research aid. Always verify direct quotes against the original recording and follow your publication's accuracy and consent policies.

Is the audio stored?

No. Audio is processed in memory for the transcription request and is not written to object storage. The transcript stays in your browser unless you export it.