FlowSTT is now in public beta for Windows and macOS. Download the latest release →
flowstt-video.mp4
FlowSTT is a free, privacy-first speech-to-text application that runs entirely on your local machine. No subscriptions, no signups, no cloud services —- your voice data never leaves your computer.
- Local transcription — Offline Whisper inference via whisper.cpp
- Hardware accelerated — CUDA on Windows, Metal on macOS
- Real-time visualization — Waveform, spectrogram, and speech activity graphs
- Multi-source audio — Microphone, system audio, or mixed mode with echo cancellation (WebRTC AEC3)
- Cross-platform — Windows (WASAPI) and macOS (CoreAudio)
- Scriptable CLI — Full command-line interface with JSON output
Download the latest release from the Releases page.
Current release: v0.3.1
Download packages
- Windows: FlowSTT_0.3.1_x64_en-US.msi
- macOS (Apple Silicon M-Series): FlowSTT_0.3.1_aarch64.dmg
- macOS (Legacy Intel x64): FlowSTT_0.3.1_x64.dmg
Linux Support Coming Soon!
MIT - Copyright 2026 KMX Consulting LLC