Agent-Reach
#77 in Agenten-FrameworksUnknown · 2× · tolest 29. Juni 2026
3
Momentum
Agent-Reach is a platform for deploying AI agents across multiple messaging and social media channels from a single codebase. The product enables agents to run simultaneously on Telegram, Discord, Slack, WhatsApp, Web, and CLI, as well as to read and search Twitter, Reddit, YouTube, and GitHub.
Momentum-Verloop
19.05.17.08.
Features
| Autonomy Level | Agent autonomously invokes tools via CLI/MCP; Agent-Reach itself is an installer/router, not execution logic – 'NOT a wrapper, agents call upstream tools directly' |
| Channels | 16+ platforms: Twitter/X, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu, Douyin, Weibo, WeChat Articles, Xiaoyuzhou Podcast, LinkedIn, Instagram, V2EX, RSS, Exa web search, arbitrary URLs |
| License | MIT license, open source (25.8K GitHub stars) |
| Modalities | Text (posts, comments, threads), video metadata/subtitles (YouTube, Bilibili), code/repo data (GitHub), web pages, RSS |
| Platform | Python CLI, runs locally (macOS/Linux/Windows); installed via pip/GitHub, usable with Claude Code, Cursor, OpenClaw, Codex, Windsurf and others |
| Price | Free / open source; optional ~$1/month residential proxy (e.g. Reddit/Bilibili from a server) |
| Release Date | Current version 1.5.0 (as of ongoing 2026 releases) |
| Tool/MCP Integrations | MCP compatible (stdio server); uses yt-dlp, gh CLI, twitter-cli, Jina Reader, Exa via mcporter, feedparser, among others |