ChatIRC — IRC-style CLI ChatGPT frontend

ChatIRC is a small curses-based IRC-like client that uses OpenAI chat models as the assistant backend. It aims for a keyboard-first, compact UX inspired by classic IRC clients like BitchX and iirc.

Highlights

• Per-room chat logs and message history
• Slash commands: /room, /rooms, /save, /load, /nick, /topic, /help, /quit, /getkey, /invite
• Tab-completion for commands/rooms and persistent input history
• Unicode-aware wrapping and optional wcwidth dependency for accurate layout
• Background API calls so the UI stays responsive

Quick start

1. Create and activate a virtualenv (recommended):

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

echo "sk-..." > openapi_key 2. Provide your OpenAI API key. Recommended: create a per-user hidden file under ~/.chatirc/openapi_key (preferred), or put openapi_key in the project directory:

Preferred (per-user):

mkdir -p ~/.chatirc
echo "sk-..." > ~/.chatirc/openapi_key
chmod 600 ~/.chatirc/openapi_key

Fallback (project-local):

echo "sk-..." > openapi_key
chmod 600 openapi_key

You can also set OPENAI_API_KEY as an environment variable.

3. Run the client:

python3 chatirc.py

Useful commands

• /getkey — opens the OpenAI API keys page in your browser
• /invite <email> — opens your mail client with a templated invite (replace the placeholder link with your GitHub release URL)
• /help — lists commands

Packaging

• To build a single-file binary using PyInstaller:

./build.sh

• A Makefile target make build runs the script. The resulting binary will appear under dist/.

Security & privacy

• Do NOT commit your openapi_key file into version control. It's ignored via .gitignore.
• Consider storing the key under ~/.chatirc/openapi_key for multi-project installs; the current default matches the working directory file openapi_key.

Contributing

Pull requests welcome. See TODO.md for in-progress ideas.

License

MIT

Assets and packaging notes

• The assets/ directory contains appicon.png and theme images. When building with PyInstaller the workflow bundles these files with --add-data so the runtime can access them. At runtime the helper get_asset_path(name) resolves the correct absolute path whether running from source or from a PyInstaller one-file bundle (via sys._MEIPASS).

• Windows: the CI attempts to convert assets/appicon.png to a .ico using ImageMagick if available and passes it to PyInstaller via --icon.

• If you want guaranteed OS-native icons ensure .ico (for Windows) and .icns (for macOS) files are present in assets/ and update the workflow to use them.

Testing

• Tests are run with pytest in CI. Locally you can run:

python3 -m pip install -r requirements.txt
pytest -q