Hey everyone! I'm the maker of RayClaw 👋 I built RayClaw because I was frustrated running separate AI bots for each chat platform. Every time I wanted the same agent on Telegram AND Discord AND Slack, I ended up maintaining three different codebases with duplicated logic. RayClaw solves this with a single Rust binary (~15MB) that connects to Telegram, Discord, Slack, Feishu/Lark, and a built-in Web UI — all sharing the same agent engine, tool system, and conversation memory. Why Rust? - Memory-safe with zero runtime overhead — no GC pauses mid-conversation - Single static binary, no Python/Node dependency hell - Small footprint: runs comfortably on a $5 VPS What makes it different: - Unified tool-calling loop across all platforms - Built-in tools: shell, file ops, web search, browser, scheduling - SQLite-based session persistence — pick up conversations where you left off - Context compaction so long chats don't blow up your token budget - Works with Anthropic and any OpenAI-compatible endpoint One config file, cargo build, done. Would love your feedback — especially if you're running AI agents in production and hitting the same multi-platform pain. Happy to answer any questions!

RayClaw is a multi-channel agentic AI - One small, safe Rust version Claw to power all your AI chats

Replies