Leon

Leon is an open-source personal assistant designed to live on your own server rather than in someone else's cloud. Built on top of Node.js, Python, and core AI concepts, it is one of the longest-running open-source personal-assistant projects — under active development since 2017 — and has attracted a large community of self-hosters who want a Siri- or Alexa-like helper they fully control. At its core, Leon does things when you ask him to. You can talk to him and he can talk back, with built-in speech-to-text and text-to-speech that can run on-device for privacy. Interaction is conversational, and Leon is organized around a modular skills system: each skill is a self-contained package that gives Leon a new controlled action or capability, from checking the weather to controlling smart-home devices. Crucially, Leon can use real tools to get work done rather than only running code in a REPL, and it keeps a layered memory so durable preferences and context persist across conversations. Leon supports both local and remote AI providers, which matters for privacy-focused users who want to keep everything on their own hardware, as well as those happy to tap cloud LLMs. The project is currently mid-way through a major rewrite on the 'Road to 2.0', which adds an agentic loop (understand a goal, choose how to handle it, use tools, remember useful information, and recover from errors), advanced memory, proactive behaviors, and local-LLM support. Leon is best understood as a project for tinkerers, self-hosters, and privacy advocates who want a personal assistant they can audit, extend, and run entirely on their own infrastructure. It is less polished and less feature-complete than commercial voice assistants, but it offers something they cannot: full ownership and control.