Linux Speech Flow

Speech-to-text for Linux. Hold a key, speak, release — text appears where you're typing.

Features

• Instant transcription — press a hotkey, speak, release — text pastes where you're typing
• Groq Whisper API — fast, accurate speech-to-text (whisper-large-v3-turbo)
• LLM post-processing — grammar cleanup and filler word removal
• Conversation mode — long-form recording with AI analysis and Q&A
• Multi-provider AI — Groq, Grok by xAI, Google Gemini
• Slack integration — record huddles, post transcriptions to Slack channels
• System tray — always accessible, minimal footprint
• Hotkey customization — remap all hotkeys from Settings
• Pipeline history — last 20 runs with SQLite storage

System Requirements

• Ubuntu 22.04+ / Debian 12 / Pop!_OS 22.04 (or compatible)
• Python 3.10+
• GTK 4 (libgtk-4-1)
• PulseAudio or PipeWire
• X11 or Wayland with XWayland
• xdotool (for text injection on X11)

Installation

From .deb package (recommended)

Download the latest .deb from Releases:

sudo apt install ./linux-speech-flow_*.deb

From PyPI

pip install linux-speech-flow

System dependencies must be installed separately (the .deb pulls these in automatically; the pip path does not):

sudo apt install python3-gi gir1.2-gtk-4.0 libgirepository1.0-dev \
  libpulse-dev xdotool xclip libnotify-bin

Quick Start

1. Run linux-speech-flow
2. The setup wizard opens on first launch
3. Enter your Groq API key (get one free)
4. Select your microphone
5. Press the hotkey (default: Ctrl+Alt+R) to start recording — release to transcribe

Autostart at login is enabled automatically — the app writes ~/.config/autostart/linux-speech-flow.desktop on first run, so it starts with your desktop session. Remove that file to disable it.

Privacy

Audio is processed by Groq's Whisper API. Conversation transcripts sent for AI analysis go to your chosen provider (Groq, Grok by xAI, or Google Gemini). No audio or transcription data is stored by the app beyond what you save locally. Your API keys are stored in ~/.config/linux-speech-flow/config.json (mode 0600, readable only by you).

API Keys

Service	Required for	Get key
Groq	Transcription (required)	console.groq.com
Grok by xAI	Conversation AI (optional)	console.x.ai
Google Gemini	Conversation AI (optional)	aistudio.google.com
Slack	Slack integration (optional)	See Slack setup guide below

Configuration

Config file: ~/.config/linux-speech-flow/config.json (created on first run)

See config.example.json for all available options with placeholder values.

Slack Setup

Slack integration requires creating a Slack App with a bot token (xoxb-) and an app-level token (xapp-, for Socket Mode). See the Slack setup guide for step-by-step instructions.

Contributing

See CONTRIBUTING.md for development setup and guidelines.

License

MIT — see LICENSE.