Documentation

Download the latest ISO from the Download page. Boot it via QEMU or write it to a USB drive and boot from UEFI. The live environment starts automatically. Run `inferra-install` to begin the guided installation.

Minimum 16 GB disk space. Inferra uses Btrfs by default with zstd compression. A swap partition is created automatically if you have more than 8 GB of disk space.

Currently Inferra uses the full disk. Dual-boot support is planned for a future release.

The configuration file is at `/etc/inferra/config.toml`. It includes UI colors, service settings, API keys for OpenRouter.ai, and hardware parameters. Changes take effect after reboot or service restart.

Yes. Set `openrouter_base_url` in `/etc/inferra/config.toml` to any OpenAI-compatible API endpoint. The NLP service will route through it.

Each capability (file management, process control, networking, etc.) runs as an independent Rust service. They communicate via gRPC through the orchestrator, which handles routing and aggregation.

Aura is the display compositor. It manages the framebuffer via KMS/DRM directly, renders the UI with wgpu (Vulkan/GL), and hosts the Wayland compositor for application windows. It runs as PID 1 in the display session.

User input is first classified by a keyword-based intent system that maps to specific actions (file read, package install, etc.). If confidence is low, it falls through to an LLM (OpenRouter.ai) for general chat. Each intent has a dedicated routing rule with its own system prompt.