Skip to content
Conduit

skill

xiaoyuzhou-podcast

Xiaoyuzhou podcast fetcher & transcriber — download, transcribe with faster-whisper, summarize

rrrrrredy/xiaoyuzhou-podcastcommunitylistedunreviewed
Use skill

Set up

Skills are installed by placing their folder where your agent loads skills.

Download ZIP from GitHub Open the repository

The download comes from github.com/rrrrrredy/xiaoyuzhou-podcast itself, not from our server, so what you get is exactly what its author published.

Install instructions come from the README

Skills do not use an MCP config. Follow the steps in the README below.

README

synced Oct 26, 2026

xiaoyuzhou-podcast

Xiaoyuzhou (小宇宙) podcast content fetcher, transcriber & summarizer.

OpenClaw Skill — works with OpenClaw AI agents

What It Does

Fetches public podcast metadata from Xiaoyuzhou (xiaoyuzhoufm.com) via the iTunes API → RSS feed pipeline, temporarily retrieves audio for local faster-whisper transcription, and generates structured summaries with key topics, quotes, and timestamps. No browser scraping needed — uses RSS instead of relying on SPA rendering.

Quick Start

openclaw skill install xiaoyuzhou-podcast
# Or:
git clone https://github.com/rrrrrredy/xiaoyuzhou-podcast.git ~/.openclaw/skills/xiaoyuzhou-podcast

One-command fetch + temporary audio cache:

bash scripts/fetch_episode.sh "忽左忽右"

Then transcribe:

python3 scripts/transcribe.py /tmp/podcast_episode.m4a /tmp/podcast_transcript.txt

Features

  • RSS-based pipeline: iTunes Search API → RSS feed → temporary audio cache (avoids SPA rendering dependency)
  • Local transcription: faster-whisper with timestamps, no external API needed
  • Structured summaries: topics, key points, keywords, highlight quotes with timestamps
  • One-click scripts: fetch_episode.sh handles search → RSS fetch → temporary audio cache in one command
  • Proxy-aware: auto-routes RSS and audio fetches through upstream proxy
  • Hard stop protection: auto-stops after 3 failures per step

Usage

Trigger with natural language:

  • "转写忽左忽右最新一期"
  • "小宇宙播客转写"
  • "播客总结" / "获取播客内容"
  • "转写这期播客"

Full pipeline:

  1. Search podcast → get RSS URL (iTunes API, no proxy needed)
  2. Download RSS XML (requires proxy)
  3. Parse latest episode metadata
  4. Temporarily fetch audio file (requires proxy, ~50-200MB)
  5. Transcribe with faster-whisper (~15-30 min for 80 min audio)
  6. Generate structured summary via LLM

Project Structure

xiaoyuzhou-podcast/
├── SKILL.md                      # Skill definition and full workflow
├── scripts/
│   ├── fetch_episode.sh          # One-click: search → temporary audio cache
│   ├── setup.sh                  # Dependency setup
│   └── transcribe.py             # Local faster-whisper transcription
├── references/
│   └── troubleshooting.md        # Debugging guide and known issues
├── README.md
├── LICENSE
└── .gitignore

Requirements

  • OpenClaw agent runtime
  • Python 3 with faster-whisper (pip install faster-whisper)
  • curl for RSS/audio fetching
  • Upstream proxy for RSS feed and audio access
  • ~200MB+ temporary disk space for audio cache; delete audio after transcript unless the user asks to keep it

License

MIT

↑↓ navigate openesc close