# Lumen

A downloadable, self-hosted AI agent framework that works immediately after installing, with a modular system of personalities, connectors, and providers that can be reconfigured without writing code.

Lumen is a downloadable AI agent framework, installed with `pip install enlumen` and started with `lumen run`, that gives a working assistant immediately rather than requiring configuration before first use. From there it can be reshaped by picking a personality, installing a module for a new capability or integration, or loading a custom `module.yaml`, without needing to write code for everyday use. The project describes itself as "WordPress, but for AI agents" -- a framework you own and run, not a SaaS or a fixed chatbot.

The lightweight Core install profile uses a built-in OpenAI-compatible HTTP client for headless deployments against OpenAI, DeepSeek, Ollama, OpenRouter, or another compatible endpoint; a `full` install profile adds LiteLLM for providers such as Anthropic, Bedrock, or Gemini. A first-run setup wizard offers a quick start with a free OpenRouter model, a personality catalog, or a custom module upload, opening into a browser chat interface with modules, memory, and settings panels.

`lumen run` targets local, personal use; `lumen server` targets a VPS or always-on machine that should stay reachable over the network.

Website: https://github.com/gabogabucho/lumen-agent