How to Use Agent-Reach
Give your AI agent eyes to see the entire internet - 13 platforms, one CLI, zero API fees. Read Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu, and more without paying for expensive API keys.
What is Agent-Reach?
Agent-Reach is a capability layer for AI agents. It is not another tool - it is an installer, health checker, router, and configuration manager for existing open-source tools (yt-dlp, gh CLI, twitter-cli, bili-cli, OpenCLI, and more).
Quick Install
pipx install https://github.com/Panniantong/agent-reach/archive/main.zip agent-reach install --env=auto
After install, run agent-reach doctor to see what works.
Platform Guide
Any Web Page
Zero configcurl -s "https://r.jina.ai/URL" // Clean markdown output
Just paste any URL. Works on articles, docs, news - anything publicly accessible.
YouTube
Zero configyt-dlp --write-sub --skip-download -o "/tmp/%(id)s" "https://youtube.com/watch?v=VIDEO_ID"
cat /tmp/VIDEO_ID.en.vtt // Read the subtitles
Get subtitles from any video. Perfect for summarizing tutorials, talks, or lectures.
Web Search
Zero configmcporter call 'exa.web_search_exa(query: "your search query", numResults: 5)'
AI-powered semantic search. Better than Google for technical questions. Free, no API key needed.
RSS Feeds
Zero configpython3 -c "import feedparser; [print(f'{e.title} - {e.link}') for e in feedparser.parse('FEED_URL').entries[:5]]"Subscribe to any RSS/Atom feed - blogs, news sites, podcasts.
V2EX
Zero configcurl -s "https://www.v2ex.com/api/topics/hot.json" -H "User-Agent: agent-reach/1.0"
curl -s "https://www.v2ex.com/api/topics/show.json?node_name=python" -H "User-Agent: agent-reach/1.0"
Popular tech community in China. Public API, no auth needed.
Bilibili
Zero configbili search "AI agents" --type video -n 5
bili hot -n 10
bili video BV1xx411c7mD
Search videos, see trending, get video details - all without login.
GitHub
Install gh CLIgh search repos "topic" --sort stars --limit 10
gh repo view owner/repo
gh issue list -R owner/repo --state open
Search repos, inspect code, check issues. Falls back to Jina Reader if gh CLI not installed.
Twitter/X
Needs cookies (1 min)twitter feed -n 20
twitter tweet URL_OR_ID
twitter user-posts @username -n 20
Read tweets, search, browse timelines. Needs cookies exported from browser once.
opencli reddit search "query" -f yaml
opencli reddit subreddit SubredditName -f yaml
opencli reddit hot -f yaml
Search posts, browse subreddits, read comments. Uses your browser login on desktop.
XiaoHongShu
Login requiredopencli xiaohongshu search "query" -f yaml
opencli xiaohongshu note NOTE_URL -f yaml
opencli xiaohongshu feed -f yaml
Search and read Xiaohongshu notes and reviews. Desktop uses browser login; server needs QR scan.
curl -s "https://r.jina.ai/https://linkedin.com/in/username"
curl -s "https://r.jina.ai/https://linkedin.com/company/name"
Read public profiles and company pages. For deeper access, install linkedin-scraper MCP.
Podcasts
Free API keyagent-reach transcribe "URL" --polish
agent-reach transcribe "URL"
Transcribe any podcast episode to text. Optional polish mode (Llama 3.3 70B) for clean summaries.
Ready to Give Your Agent Eyes?
Star the repo, install it, and let your AI agent browse the entire internet - no API fees.