Skip to content

flipfile v0.8.0

Choose a tag to compare

@github-actions github-actions released this 12 Nov 09:08

✨ flipfile v0.8.0 — Transcribe

flipfile v0.8.0 writes down what is said. Drop a lecture, a meeting, a WhatsApp voice note or a video, and get subtitles or a transcript in the language it was spoken in, right on your computer.


🎙️ Transcribe

  • Where: Music → Transcribe, or Video → Tools → Subtitles.
  • Subtitles: SRT or VTT, named after the video (lecture.mp4 → lecture.srt) so players pick them up straight away. One sentence per subtitle, long ones split after a comma, never more than two lines.
  • Transcripts: TXT, Word or Markdown, in paragraphs where the speaker paused, with a timestamp at each one.
  • Any recording: MP3, M4A, WAV, FLAC, OGG and Opus voice notes, and the sound of MP4, MOV, MKV and WebM videos, with nothing to install.
  • 99 languages, detected automatically (Indonesian, English, Javanese, Malay, Japanese…).

✦ Then the gist (Mac)

  • Drop a recording or a video on AI ✦ to have it transcribed and summarised in one go: the decisions, to-dos and deadlines of an hour-long meeting.

⚡ On your computer

  • On a Mac with Apple silicon it runs on the graphics card: about 15× faster than real time, so an hour-long lecture takes about 4 minutes. On other computers it runs on the processor, more slowly, and the circle shows the progress.
  • Private and offline: OpenAI's Whisper runs on your computer, and nothing is uploaded.

🔒 Verify your download

  • Checksums: SHA256SUMS.txt below lists the SHA-256 of every file. Check with shasum -a 256 -c SHA256SUMS.txt --ignore-missing (Windows: certutil -hashfile <file> SHA256).
  • Build provenance: every installer was built by GitHub Actions from the v0.8.0 tag and carries a signed attestation: gh attestation verify <file> --repo wcksnrdn/flipfile. More in SECURITY.md.

💡 The speech model adds about 190 MB to the download.

💡 flipfile now needs macOS 10.15 Catalina or later (as Tauri 2 always did).

💡 Already on v0.7.x? Use Check for Updates… from the menu bar / tray.

Speech model: Whisper by OpenAI (MIT), run by whisper.cpp (MIT). Full Changelog: CHANGELOG.md · Compare v0.7.1...v0.8.0