A self-hosted, evolving personal planner where a Claude Code agent on your Mac rewrites the app from chat requests and ships updates to your devices.
- Stars
- 269
- Forks
- 39
- Open issues
- 2
- Last commit
- 2 days ago
What is it?
Xiryn is an open-source personal agent that runs entirely on your own computer. Each agent gets a sandboxed Linux desktop in Docker with a shell, persistent files and Chromium, so it can browse the web, write and run code, and work in the background without touching your own mouse, screen or files. You can watch the agent's desktop live and take over at any point.
Agents work with any configured model provider, including OpenAI, Anthropic, Google, Ollama and LM Studio, using your own API keys. Conversations, memory and settings are stored in a single SQLite file on your disk. Saved logins and payment details sit in an encrypted vault whose key is held by the operating system, and agents only see masked values. Every payment requires your approval.
The project includes scheduled routines, multi-agent squads that delegate tasks to each other, usage and cost tracking with an optional monthly budget, and Telegram pairing for remote control and hand-off notifications. A one-line setup script installs Git, Node.js, Docker and the agent desktop image.
Key features
- 01Each agent runs in its own Docker Linux desktop with a shell and Chromium, separate from your system
- 02Works with 11 model providers including local Ollama and LM Studio, using your own keys
- 03Encrypted vault for logins and cards, with per-payment approval before any purchase
- 04Telegram pairing for remote control, live desktop links and completion notices
- 05Scheduled routines and squads of agents that delegate work to one another
- 06Per-reply cost tracking and optional monthly budget for paid models
Use cases
- Running research, browsing or coding tasks in the background while you keep working on your laptop
- Keeping conversations and agent memory private on your own disk instead of a vendor's cloud
- Using free local models through Ollama for an agent that never sends prompts off your machine
- Controlling a home-hosted agent from your phone to handle recurring tasks like morning briefs
Topics
FAQ
- What is xiryn?
- A self-hosted personal AI agent where each agent runs in its own Linux desktop container on your computer, with any model, an encrypted vault and Telegram control. Xiryn is an open-source personal agent that runs entirely on your own computer. Each agent gets a sandboxed Linux desktop in Docker with a shell, persistent files and Chromium, so it can browse the web, write and run code, and work in the background without touching your own mouse, screen or files. You can watch the agent's desktop live and take over at any point.
- Is xiryn free to use? What license does it have?
- xiryn is open source under the MIT license. Check the license terms before using it in commercial products.
- Is xiryn still maintained?
- Actively maintained. The last commit to xiryn was on Oct 20, 2026.
- What can I use xiryn for?
- Running research, browsing or coding tasks in the background while you keep working on your laptop Keeping conversations and agent memory private on your own disk instead of a vendor's cloud Using free local models through Ollama for an agent that never sends prompts off your machine Controlling a home-hosted agent from your phone to handle recurring tasks like morning briefs
- Where can I get xiryn?
- The source code and installation instructions are on GitHub: https://github.com/SouravBeraAkaSpeed/xiryn
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open-source LLM gateway that routes requests from AI agents and apps to many model providers through one OpenAI-compatible endpoint.
Local-first personal OS agent built on Hermes that logs mood, sleep, meals and habits, then surfaces patterns across them over time.
A Telegram-driven personal agent that turns voice notes and messages into typed, linked markdown cards in your Obsidian vault, running on your own server.
A personal AI assistant that runs on macOS, Windows, iOS, Telegram, Slack and more, with persistent memory and proactive follow-ups.
A self-hosted personal AI agent server in a single Rust binary, with sandboxed tool execution, multiple LLM providers, and chat channels like Telegram and Discord.