Lightweight AI AgentsSpeaks to
Discord
Deploy AI agents that communicate across Email, Telegram, Discord, Slack, and custom channels from a single lightweight Rust runtime.
Chat with your agents straight from the terminal. Launch the interactive TUI for a fast, keyboard-driven session — no browser required, fully self-hosted.
Everything Your Agents Need.
From channels and memory to security and model routing.
Lightweight by Design
A compact Rust-powered runtime built for fast startup, low memory use, and predictable production behavior.
~12 MB
Binary
~200ms
Cold Start
~15 MB
Idle Memory
No GC
Predictable
Extend with Clawhub
Add reusable skills, MCP integrations, and ready-to-use modules from Clawhub.
Web UI
Manage and interact with your agents through an intuitive interface.
18 Channels
iMessage, Signal, Discord, Telegram, DingTalk, and more — reach your agents wherever the conversation happens.
30+ LLM Providers
Route every agent to Claude, GPT, Gemini, DeepSeek, Ollama, and more — swap models without changing your setup.
Bring Your Knowledge
RantAIClaw supports RAG (Retrieval-Augmented Generation) out of the box — turn PDFs, code, docs, images, and web pages into knowledge your agents retrieve when answering.
One command.
No compiler.
# Auto-detects your platform, downloads the matching prebuilt binary from the latest release
$ curl -fsSL https://raw.githubusercontent.com/RantAI-dev/RantAIClaw/main/scripts/bootstrap.sh | bash
Prefer a Linux toolchain on Windows? Install WSL2 and run the Linux one-liner inside the Ubuntu shell.
Read the docs. Ship an agent.
[where agents live and work]