Self-hostable multi-agent personal AI assistant built on LangGraph, connecting mail, calendar, documents and tasks with human approval before actions run.
opencrust-org/opencrust
A secure, lightweight Rust-based personal AI agent framework running across chat channels like Telegram, Discord, Slack and WhatsApp.
- Stars
- 148
- Forks
- 28
- Open issues
- 39
- Last commit
- 19 hours ago
What is it?
OpenCrust is a single-binary personal AI assistant platform written in Rust. It runs LLM-backed agents that answer messages arriving on Telegram, Discord, Slack, WhatsApp, WhatsApp Web, LINE, WeChat, iMessage and MQTT, and it includes a built-in web chat and terminal chat client. Credentials are stored in an AES-256-GCM encrypted vault, and the gateway requires pairing by default and binds to localhost.
The agent side covers 15 LLM providers including Anthropic, OpenAI, Ollama and several OpenAI-compatible endpoints. It supports MCP servers over stdio and HTTP, document RAG with hybrid vector and keyword search over SQLite, named multi-agent routing with a handoff tool, cron and one-shot scheduling, and voice input and output through Whisper and TTS engines.
Skills are defined as SKILL.md files and can be installed from agentskills.io-compatible URLs, with a prompt-injection scan before installation. The project also includes config hot-reload, per-user rate limits and token budgets, and an `opencrust migrate openclaw` command for importing an existing OpenClaw setup.
Key features
- 01Runs AI agents across Telegram, Discord, Slack, WhatsApp, LINE, WeChat, iMessage and MQTT channels.
- 02Stores API keys and tokens in an AES-256-GCM encrypted vault.
- 03Connects to MCP servers over stdio and Streamable HTTP.
- 04Answers questions using ingested documents through hybrid vector and keyword search.
- 05Routes tasks between named agents with a built-in handoff tool.
- 06Loads agent skills from SKILL.md files and scans them before installation.
Use cases
- Running a self-hosted personal assistant that answers you on Telegram or Discord.
- Giving an LLM agent access to your documents for question answering with citations.
- Connecting a local Ollama model to chat channels without a cloud dependency.
- Migrating an existing OpenClaw setup to a Rust-based runtime.
Quick start
- 1
Install on Linux or macOS via the official install script
- 2
Run interactive setup to pick an LLM provider and channels
opencrust init - 3
Start the agent gateway
opencrust start - 4
Diagnose configuration, connectivity and database health
opencrust doctor
Things to know
- Installation relies on a curl-pipe-to-shell install script, and building from source needs Rust 1.85+.
- Several channels and features (iMessage, WhatsApp Web, embeddings, WASM plugins) have platform or setup requirements.
In collections
Topics
FAQ
- What is opencrust?
- A secure, lightweight Rust-based personal AI agent framework running across chat channels like Telegram, Discord, Slack and WhatsApp. OpenCrust is a single-binary personal AI assistant platform written in Rust. It runs LLM-backed agents that answer messages arriving on Telegram, Discord, Slack, WhatsApp, WhatsApp Web, LINE, WeChat, iMessage and MQTT, and it includes a built-in web chat and terminal chat client. Credentials are stored in an AES-256-GCM encrypted vault, and the gateway requires pairing by default and binds to localhost.
- Is opencrust free to use? What license does it have?
- opencrust is open source under the MIT license. Check the license terms before using it in commercial products.
- Is opencrust still maintained?
- Actively maintained. The last commit to opencrust was on Nov 7, 2026.
- What can I use opencrust for?
- Running a self-hosted personal assistant that answers you on Telegram or Discord. Giving an LLM agent access to your documents for question answering with citations. Connecting a local Ollama model to chat channels without a cloud dependency. Migrating an existing OpenClaw setup to a Rust-based runtime.
- How do I install opencrust?
- Install on Linux or macOS via the official install script Run interactive setup to pick an LLM provider and channels: opencrust init Start the agent gateway: opencrust start Diagnose configuration, connectivity and database health: opencrust doctor
- Where can I get opencrust?
- The source code and installation instructions are on GitHub: https://github.com/opencrust-org/opencrust
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open-source personal AI assistant and Agent Harness that plans tasks, runs tools and skills, and keeps memory across web and IM channels.
A self-hosted AI assistant for households and small teams, with multiple experts, IM channel integrations, and a web dashboard, all in one process.
Open-source AI personal assistant for planning, writing and personal knowledge, runnable on a Mac or self-hosted with your own model keys.
A self-hosted personal AI assistant with visible memory and a Pi coding-agent workspace, reachable from browsers, phones and WeChat.
A self-hosted personal AI assistant that runs on your machine or in the cloud and connects to chat apps like DingTalk, Lark, Discord and Telegram.