Quick start
Install and configure a minimal provider path, then verify one transport at a time.
Install and configure
Use Python 3.11 or newer with dependency extras appropriate to the host. The package declares mushegh and mushegh-setup entry points. Run mushegh-setup for the interactive configuration workflow.
Keep generated secrets private. Begin with one brain and one known audio or text client before enabling optional accounts or unattended work.
Add the HUD
From clients/hud-spatial install with npm ci and run npm run dev for loopback development, or npm run build for gateway static serving. Configure authenticated brain and private gateway connections server-side.
Verify a real request, audible response, fresh source record and camera frame separately. Use synthetic approvals for testing, never an unknown real pending task.
Minimal installation commands
From a cloned checkout, create a virtual environment and install with: python -m venv .venv; then use that environment's Python to run python -m pip install -e ".[hud]". The hud extra supplies read and gateway dependencies; desktop voice requires additional extras declared in pyproject.toml.
Run mushegh-setup from the activated environment, then python -m mushegh.brain.server for the brain. Build the HUD from clients/hud-spatial using npm ci and npm run build before starting the private gateway. Keep these processes separate when diagnosing failures.
Implementation sources
Repository paths below describe the implementation. They do not certify a running deployment or a configured external account.
pyproject.tomlsrc/mushegh/setup_wizard.pysrc/mushegh/cli.py.env.example
Subsystem directory · HUD guide · Verification and known limits