Leon
Open source personal assistant you run on your own server. Talk to it via text or voice, extend it with custom skills written in Python or Node.js, and keep all interactions on your own infrastructure
Quick Start
git clone https://github.com/leon-ai/leon && cd leon && npm install && npm run setup && npm start Overview
Leon is an open source personal assistant that runs on your own server. You interact with it via text or voice, and it responds through a browser-based interface or an API. Unlike commercial assistants, Leon keeps all interactions on infrastructure you control — no audio clips or query history sent to Google, Amazon, or Apple.
The assistant is built around a skill architecture. Each skill is a module — written in Python or Node.js — that handles a category of intent. Leon ships with skills for weather, timers, translations, news, and a handful of other basic functions. The real extensibility is in writing your own: a skill that queries your own APIs, controls a local service, fetches data from your homelab, or automates a personal workflow that no commercial assistant would support.
Speech-to-text and text-to-speech are pluggable. Leon supports multiple STT backends, including options that run fully offline, and multiple TTS engines for voice output. Setting up a fully offline voice pipeline requires configuration work, but the architecture supports it.
The honest comparison to make is against newer AI-backed alternatives. Leon predates the LLM wave and its default skill system is rule-based and intentional, not generative. It cannot answer general knowledge questions or handle arbitrary tasks the way a Claude or GPT-backed assistant can. For users who want a privacy-preserving assistant they can extend with custom automations, Leon is a structured, well-maintained framework. For users who want broad general-purpose capability out of the box, newer self-hosted options like OpenClaw provide a different kind of answer.
With 17.3k GitHub stars and an MIT licence, Leon has a genuine community behind it and an active roadmap.
Leon: Pros & Cons
| Pros (The Wins) | Cons (The Friction) |
|---|---|
| 17.3k stars: Active project with an established community. | Limited built-in skills: Useful capabilities require writing your own modules. |
| Extensible skills: Custom Python or Node.js modules for your use cases. | Not LLM-backed by default: No general-purpose Q&A; intent-based, not generative. |
| Privacy first: All interactions stay on your own server. | Technical setup required: Node.js + Python; voice pipeline needs configuration. |
| MIT, no cloud lock-in: No dependency on any commercial AI provider. | Voice quality varies: Depends on the STT backend you configure. |
Use Cases
Specific ways to use Leon for your workflow.
Deployment Strategy
Recommended ways to host Leon in your own environment.