Leon is a free, open-source personal AI assistant you can self-host. Explore its voice/text features, skills system, architecture, and alternatives.
Category
Developer Tools
Pricing
Free / Open Source, from Free
Verified
Not yet
Last updated
July 18, 2026
Founded
2017
Free PlanWindowsOpen SourceAISelf-HostedMac
Overview
Leon (getleon.ai) is a free, open-source, self-hosted personal AI assistant created by developer Louis Grenard, with roots dating to 2017 and active development since 2019. Released under the MIT License, it supports both voice and text interaction and can run fully offline for maximum privacy.
The project has grown to roughly 17,400 GitHub stars and 1,500 forks, with sponsorship from OpenAI, JetBrains, and MacStadium, and is currently undergoing a major architectural evolution toward agentic, tool-based execution in its '2.0 Developer Preview' phase.
Key Features
Leon supports three execution modes: 'smart' (automatic strategy selection), 'controlled' (deterministic native skills), and 'agent' (step-by-step planning for multi-step tasks), backed by a layered memory system for preferences, daily context, and discussion history.
Its extensible skills architecture splits into native skills (fast, built-in) and agent skills (flexible, tool-based), with Node.js and Python bridge layers letting developers build and share custom capabilities.
Pricing
Leon is completely free and open source under the MIT License, with no paid tiers or subscription fees for the software itself.
Users self-host Leon on their own server (Linux, macOS, or Windows), so the only costs involved are the user's own infrastructure and, optionally, any third-party cloud AI provider fees if configured to use remote models.
Key Features
Voice and text interaction — Supports both spoken conversation via integrated text-to-speech/speech-to-text and conventional text-based chat interaction.
Offline-capable, privacy-focused design — Can be configured to run entirely offline, removing dependence on third-party cloud services for core functionality.
Multiple execution modes — Offers 'smart,' 'controlled,' and 'agent' modes for automatic strategy selection, deterministic skills, or multi-step agentic planning.
Layered memory system — Tracks user preferences, ongoing daily context, and historical discussion separately for more persistent, personalized behavior.
Extensible skills architecture — Supports native skills (built-in, fast) and agent skills (flexible, tool-based) that developers can create and share.
Local and remote AI provider support — Can integrate with local, on-device AI models or remote cloud-hosted providers depending on the user's privacy and capability needs.
Cross-platform self-hosting — Runs on Linux, macOS, and Windows via CLI setup, Docker, or manual installation.
Polyglot bridge architecture — Node.js and Python bridge layers with toolkit definitions allow skills to be built in the language best suited to the task.
Pros & Cons
Pros
Completely free and open source under the MIT License
Can run fully offline for strong privacy and no dependency on third-party cloud services
Long project history since 2017 with a large, active GitHub community (~17,400 stars)
Backed by notable sponsors including OpenAI, JetBrains, and MacStadium
Flexible agentic execution model for multi-step task planning, not just single-turn commands
Cons
Requires self-hosting and a nontrivial technical setup (Node.js, Python, Pipenv dependencies)
Currently mid-transition to a '2.0 Developer Preview' architecture, meaning ongoing breaking changes
No official commercial support; relies on community documentation and GitHub issues
Voice recognition and TTS quality depend on which providers are configured, offline options may be less accurate than top commercial cloud services
Not a plug-and-play consumer product like Alexa or Google Assistant
Pricing
Free / Open Source Free N/A
Frequently Asked Questions
Is Leon (getleon.ai) free?
Yes, Leon is completely free and open source under the MIT License.
Who created Leon?
Leon was created by developer Louis Grenard, with the project dating back to 2017 and active development since 2019.
Can Leon run fully offline?
Yes, Leon can be configured to operate entirely offline, without relying on third-party cloud services, for maximum privacy.
What platforms does Leon support?
Leon runs on Linux, macOS, and Windows, and can be installed via a CLI setup wizard, Docker, or manual installation.
Does Leon support voice interaction?
Yes, Leon supports both voice (via text-to-speech and speech-to-text) and text-based chat interaction.
Is Leon still in active development?
Yes, Leon is undergoing a major architectural transition and is currently in a '2.0 Developer Preview' phase with ongoing active development.