Feishu Podcast

Feishu Podcast

Workflow

Use the bundled script first:

python3 ~/.codex/skills/feishu-podcast/scripts/podcast_to_minutes.py "<podcast episode URL>" --output-dir .

The script:

1. Resolves the audio URL from common podcast pages.
   • Simplecast episode URLs via https://api.simplecast.com/episodes/search.
   • Direct .mp3 URLs.
   • RSS feeds and HTML pages with RSS or MP3 links.
2. Downloads the audio as an .mp3 file.
3. Runs lark-cli drive +upload --file <mp3> to get file_token.
4. Runs lark-cli minutes +upload --file-token <file_token> to create the Feishu Minutes link.
5. Prints JSON with mp3_path, file_token, drive_url, and minute_url.

Required Feishu Permissions

The Feishu upload path needs these scopes:

• drive:file:upload
• minutes:minutes.upload:write

If lark-cli returns missing_scope, start the exact auth command from the CLI hint, wait for the user to approve the browser/device link, then rerun the script.

Defaults

• Save the MP3 in the current working directory unless the user asks for another folder.
• Use a readable filename derived from the podcast title.
• Do not fetch transcripts, AI summaries, or edit the generated document unless the user explicitly asks after the Minutes link exists.
• If the user asks for transcript or summary next, use the generated minute_url token with lark-cli vc +notes --minute-tokens <token>.

Notes

• This is a write workflow: it uploads a file to Feishu Drive and creates a Feishu Minutes asset. Treat the user's request to "put it into Feishu Minutes" as confirmation for these two writes.
• If the podcast resolver fails, inspect the page manually for RSS feed metadata or embedded audio URLs, then patch or bypass the script with a direct MP3 URL.