PythonClaw and friendlyclaw
These are ecosystem siblings within the OpenClaw framework—PythonClaw is a Python implementation of the core autonomous agent architecture, while FriendlyClaw extends that foundation with multi-agent delegation and proactive control capabilities.
About PythonClaw
ericwang915/PythonClaw
OpenClaw reimagined in pure Python — autonomous AI agent with memory, RAG, skills, web dashboard, voice input, daemon, and multi-channel support.
Implements a hybrid RAG system combining BM25 lexical search with dense embeddings and reciprocal rank fusion, plus LLM-based re-ranking for semantic relevance. Built on provider-agnostic LLM abstraction supporting DeepSeek, Claude, Gemini and OpenAI-compatible APIs, with persistent markdown-based memory, three-tier progressive skill loading from ClawHub marketplace, and concurrent multi-channel support (CLI, web, Telegram, Discord, WhatsApp) via daemon mode with per-session isolation and concurrency locks.
About friendlyclaw
me2Doc/friendlyclaw
FriendlyClaw 🦀 — Pro-Grade Autonomous Hive Operative. Multi-agent delegation, semantic RAG memory, and proactive system control.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work