CamelAGI is a lightweight OpenClaw alternative that lets you run Claude Code through Telegram, web apps, or terminal using your own machine. It sends commands to your locally installed Claude Code, so authentication stays on your device under your own Anthropic subscription. Built on Claude Agent SDK, CamelAGI also adds self-hosted multi-agent control, memory, tools, MCP servers, and a much smaller codebase designed for easier maintenance.

Tines — Build agents & automations integrated across your workspace

Build agents & automations integrated across your workspace

I’ve used OpenClaw heavily in several products and genuinely respect what it made possible. What pushed me to build CamelAGI was maintainability: OpenClaw’s codebase is powerful, but very large and difficult to extend when you want to move fast or customize deeply. So I reverse-engineered the core ideas and rebuilt the system around a much smaller codebase (under 20k lines), while keeping most of the capabilities I personally needed. CamelAGI adds what mattered most for my own workflow: remote Claude Code access through Telegram, cleaner memory, and stronger tool orchestration. Under the hood, CamelAGI uses Claude Agent SDK as the runtime, which gave much better behavior for reasoning, tool use, and long-running agent tasks. Would genuinely love feedback from people who know OpenClaw well.

Tines — Build agents & automations integrated across your workspace

Build agents & automations integrated across your workspace

CamelAGI

Run self-hosted AI agents, alternative to OpenClaw

Run self-hosted AI agents, alternative to OpenClaw