Agent Reach

Summary

Panniantong/Agent-Reach is a CLI capability layer (not a wrapper scraper) that selects, installs, health-checks, and routes free backends so agents can read/search the web, YouTube, RSS, GitHub, Twitter/X, Bilibili, Reddit, Facebook, Instagram, XiaoHongShu, LinkedIn, V2EX, Xueqiu, and Xiaoyuzhou. Install by handing the agent docs/install.md; agent-reach doctor reports channel status. Zero-config paths for web/YouTube/RSS/public GitHub/Bilibili search; login-gated platforms use user-exported cookies or OpenCLI Chrome sessions (cookies stay local). MIT. Python >=3.10. Works with Claude Code, OpenClaw, Cursor, Windsurf. README: no affiliated token/crypto project.

Keywords

agent-reach · mcp · cli · scraping · twitter · reddit · youtube · openclaw · xiaohongshu

Topics

agent tools, web scraping, MCP

Research notes

• Primary: GitHub README + API (MIT, Python, 71695 stars / 6089 forks at check). Canonical full_name Panniantong/Agent-Reach; leftover URL from backlog keeps the trailing slash. Author display name Pnant (@Panniantong). Discord posted the repo. CLI/capability layer, not a new hosted corpus, so no datasets_local row. README warns of unaffiliated crypto tokens using the name.