OpenClaw🦞Finder

The official full-stack runtime with the clearest docs, broadest integrations, and the best baseline for understanding the ecosystem.

A compact Python branch that keeps the agent loop easy to read, learn from, and fork without dragging in the full stack.

A Rust runtime built for migration, performance, and lower-level control rather than the easiest first-time onboarding.

A Go-based lightweight branch aimed at tiny hardware, low memory, and always-on deployments that need a smaller footprint.

A container-first branch focused on safer isolation, clearer self-hosting boundaries, and more controlled deployments.

A more practical agent workstation with easier setup, local-model support, and built-in chat app integrations.

A Rust branch centered on privacy, validation, and safer execution for teams that care about trust boundaries.

A collaborative multi-agent branch focused on transparent, human-in-the-loop coordination through shared Matrix rooms.

A packaged route for WeChat and QQ usage that is easier to try, but still gated behind a beta activation code.

A single static Zig binary that delivers the full OpenClaw experience in under 1 MB of RAM, with <8 ms startup and no runtime dependencies.

An Android APK that turns a phone into an LLM-controlled device, wiring accessibility services and messaging channels into an autonomous agent loop.

A packaged OpenClaw desktop client with a bundled Node runtime, multi-provider key management, and one-minute install on macOS and Windows.

A community-maintained Chinese distribution of OpenClaw with Feishu, DingTalk, WeCom integrations, DeepSeek/Qwen defaults, and domestic mirrors.

The official marketplace and registry for discovering, installing, and publishing OpenClaw skills.

The official workflow format for chaining agents, tools, and steps into more structured OpenClaw flows.

An open-source desktop research assistant built on the OpenClaw ecosystem, with scheduled monitoring, web browsing, and multi-channel delivery.

An experimental social platform where AI agents post, comment, follow, and get studied under incentive and governance rules.

An OpenClaw workflow engine for running multi-step tool pipelines with approvals, resume tokens, and stronger auditability.

A memory framework for always-on agents, with persistent storage, structured retrieval, and a memory-as-filesystem model.

A Cloudflare-based deployment template for running OpenClaw or Moltbot with Workers, Sandbox containers, R2, and browser automation.

An OpenClaw plugin that gives agents phone-calling abilities through providers such as Twilio, Telnyx, and Plivo.

A pluggable context engine that replaces sliding-window truncation with a DAG of summarized messages, so older context can be searched and recovered on demand.

A community fork of Lossless Claw that fixes CJK token estimation and common session bugs, so Chinese, Japanese, and Korean users get accurate compaction.

A free scanner from Astrix Security that uses read-only EDR telemetry to find unmanaged OpenClaw deployments inside enterprise environments.

A Chinese agent social platform built on an OpenClaw-adapted stack, focused on practical multi-agent collaboration and local-first governance.