# Transkribeer

> Transkribeer is a free desktop app (macOS, with Windows/Linux in beta) that transcribes audio files to text using OpenAI (gpt-4o-transcribe), Google Gemini, or a fully local and offline Whisper model (mlx-whisper, Apple Silicon). The interface is in English; the app transcribes audio in any supported language.

Key facts:

- Free to download and use. Three transcription modes: Instant (zero setup, pay per second from a prepaid credit balance, currently €0.04 per minute of audio), bring-your-own OpenAI or Gemini API key, or the local Whisper mode which needs no key and no internet.
- Records audio directly in the app: microphone, system audio (e.g. the sound of a call or meeting), or both mixed into one track — no virtual audio drivers needed (uses macOS ScreenCaptureKit). Recordings are saved as m4a and can be transcribed immediately.
- Supports mp3, m4a, wav, ogg, flac, webm, aac and wma, with drag & drop of multiple files at once.
- Long recordings are automatically split at natural pauses in speech (never mid-sentence) and reassembled after transcription — no duration limits.
- Multiple providers can run side by side on the same file to compare transcription quality.
- Transcriptions are literal (verbatim): the spoken words are written down without summarizing or interpretation, and saved as a .txt file next to the audio file.
- Runs on macOS 12+ (Apple Silicon & Intel), with Windows and Linux builds in beta. FFmpeg ships inside the app — nothing to install.
- Built by Sander Versluys. Auto-updates via GitHub Releases.

## Links

- [Homepage (Dutch)](https://transkribeer.com/)
- [Homepage (English)](https://transkribeer.com/en/)
- [FAQ — how to get OpenAI and Gemini API keys](https://transkribeer.com/#faq)
- [Download latest version (macOS .dmg)](https://github.com/sanderversluys/transkribeer-releases/releases/latest)
- [Releases on GitHub](https://github.com/sanderversluys/transkribeer-releases)