x-reader

Universal content reader — fetch, transcribe, and digest content from any platform.

Give it a URL (article, video, podcast, tweet), get back structured content. Works as CLI, Python library, MCP server, or Claude Code skills.

简体中文： README.zh.md / README.zh-CN.md

What It Does

Any URL → Platform Detection → Fetch Content → Unified Output
              ↓                      ↓
         auto-detect           text: Jina Reader
         7+ platforms          video: yt-dlp subtitles
                               audio: Whisper transcription
                               API: Bilibili / RSS / Telegram

The Python layer handles text fetching and YouTube subtitle extraction. The Claude Code skills (optional) add full Whisper transcription for video/podcast and AI-powered content analysis.

Three Layers

x-reader is composable. Use the layers you need:

Layer	What	Format	Install
Python CLI/Library	Basic content fetching + unified schema	See Install	Required
Claude Code Skills	Video transcription + AI analysis	Copy skills/ to your Claude Code skills directory	Optional
MCP Server	Expose reading as MCP tools	python mcp_server.py	Optional

Layer 1: Python CLI

# Fetch any URL
x-reader https://mp.weixin.qq.com/s/abc123

# Fetch a tweet
x-reader https://x.com/elonmusk/status/123456

# Fetch multiple URLs
x-reader https://url1.com https://url2.com

# Login to a platform (one-time, for browser fallback)
x-reader login xhs

# View inbox
x-reader list

Layer 2: Claude Code Skills

Requires cloning the repo (not included in pip install).

For video/podcast transcription and content analysis:

skills/
├── video/       # YouTube/Bilibili/podcast → full transcript via Whisper
└── analyzer/    # Any content → structured analysis report

Install:

export CLAUDE_SKILLS_DIR="/path/to/claude-code-skills"
mkdir -p "$CLAUDE_SKILLS_DIR"
cp -r skills/video "$CLAUDE_SKILLS_DIR/video"
cp -r skills/analyzer "$CLAUDE_SKILLS_DIR/analyzer"

Then in Claude Code, just send a YouTube/Bilibili/podcast link — the video skill auto-triggers and produces a full transcript + summary.

Layer 3: MCP Server

Requires cloning the repo (mcp_server.py is not included in pip install).

git clone https://github.com/runesleo/x-reader.git
cd x-reader
pip install -e ".[mcp]"
python mcp_server.py

Tools exposed:

• read_url(url) — fetch any URL
• read_batch(urls) — fetch multiple URLs concurrently
• list_inbox() — view previously fetched content
• detect_platform(url) — identify platform from URL

Claude Code config (~/.claude/claude_desktop_config.json):

{
    "mcpServers": {
        "x-reader": {
            "command": "python",
            "args": ["/path/to/x-reader/mcp_server.py"]
        }
    }
}

Supported Platforms

Platform	Text Fetch	Video/Audio Transcript
YouTube	✅ Jina	✅ yt-dlp subtitles → Groq Whisper fallback
Bilibili (B站)	✅ API	✅ via Claude Code skill
X / Twitter	✅ oEmbed → FxTwitter → Article/Jina → Playwright	—
WeChat (微信公众号)	✅ Jina → Playwright	—
Xiaohongshu (小红书)	✅ Jina → Playwright*	—
Telegram	✅ Telethon	—
RSS	✅ feedparser	—
小宇宙 (Xiaoyuzhou)	—	✅ via Claude Code skill
Apple Podcasts	—	✅ via Claude Code skill
Any web page	✅ Jina fallback	—

*XHS requires a one-time login: x-reader login xhs (saves session for Playwright fallback)

X Articles and login-required X pages can use a saved local browser session: x-reader login twitter

YouTube Whisper transcription requires `GROQ_API_K

... View full README on GitHub