OpenSquawk

OpenSquawk is an open, cost-efficient AI-ATC platform for flight simulators. Built with Nuxt 4, it provides a web interface that combines training, radio communication, and classroom content for simulator pilots. On the server side, via Node.js Open AI deliver speech synthesis, decision trees, and account management. Goal: a community-driven alternative to other costly AI-ATC solutions while being open and extensible.

Current Features

• Modern web UI (Nuxt 4, Tailwind, Vuetify) with landing and classroom pages
• Server APIs for ATC logic, decision trees, classroom content, and authentication
• MongoDB storage
• OpenAI integration for speech and text models with configurable voices
• Optional local Speaches server for self-hosted speech in/output

Roadmap & Integrations

• Simulator Bridges: Planned connection to Microsoft Flight Simulator and X-Plane to receive live flight data. Implementation is scheduled but will take a few more months.
• SimBrief: Upcoming direct integration of SimBrief data (flight plans, briefings) for seamless cockpit dispatch info.
• VATSIM: No integration for now due to unclear licensing. We are monitoring the situation but will avoid coupling until approval is granted.

Requirements

Component	Note
Node.js 22	Specified in package.json.
Yarn
MongoDB 7+	Local instance or hosted database. Configure via MONGODB_URI.
ffmpeg	Needed for audio processing (fluent-ffmpeg), must be on the system PATH.
OpenAI API Key	For TTS and LLM calls (OPENAI_API_KEY).
Live ATC backend	The /pm Live ATC mode requires the Python decision backend (OpenSquawk-LiveATC-api). Point NUXT_PUBLIC_RADIO_BACKEND_URL at it (default http://127.0.0.1:8000). Other features run without it.
Optional: Piper TTS	For local speech (pip install "piper-tts[http]").

You can connect to any OpenAI-compatible (self-hosted or third-party) API by setting OPENAI_BASE_URL to the service's base URL.

Setup

1. Clone the repository:

   git clone https://github.com/OpenSquawk/OpenSquawk.git
   cd OpenSquawk
   

2. Enable corepack and install dependencies:

   corepack enable
   yarn install
   

3. Configure environment variables:

   cp .env.example .env
   

   Key variables:

   • MONGODB_URI: MongoDB connection string (default mongodb://127.0.0.1:27017/opensquawk).
   • JWT_SECRET & JWT_REFRESH_SECRET: Random strings for tokens.
   • OPENAI_BASE_URL: Optional override for any OpenAI API–compatible service; defaults to the OpenAI platform when unset.
   • OPENAI_API_KEY, optional OPENAI_PROJECT, LLM_MODEL, TTS_MODEL, VOICE_ID.
   • ATC_OUT_DIR: Directory for generated audio files.
   • FLIGHTLAB_TTS_CACHE_DIR: Disk cache directory for FlightLab TTS audio (defaults to ./.cache/flightlab-tts, git-ignored).
   • USE_PIPER, PIPER_PORT: Enable local Piper TTS instance.
   • SMTP settings (NOTIFY_*) if emails should be sent.

4. Start MongoDB and ensure ffmpeg is on the PATH.

5. Launch the dev server:

   yarn dev
   

   Interface runs at http://localhost:3000 with hot-reload.

Production & Preview

• Build:

  yarn build
  

• Local preview:

  yarn preview
  

• For custom hosting, serve the Nuxt directory (yarn start).

Optional Local Speech Services

Piper TTS

For speech output without OpenAI:

pip install "piper-tts[http]"
piper-http-server --port 5001

Set USE_PIPER=true and adjust PIPER_PORT. Model ID (SPEECH_MODEL_ID) can also be set in .env.

Local STT/TTS

Local models via WSL+Docker is planned but not yet implemented. Placeholders and TODOs exist in the code.

Project Structure

app/          # Nuxt frontend (pages, components, stores)
server/       # API routes, services, utilities (ATC TTS/STT, Auth, flow editor)
shared/       # Shared types and helpers
content/      # Markdown news & CMS content

Contributing & Community

We welcome issues, pull requests, and ideas—especially Nuxt/Node expertise, ATC knowledge, MSFS/X-Plane test flights, and infrastructure/cost optimization. Contact: info@opensquawk.de.

OpenSquawk thrives on community collaboration—priorities are set together. Join in!