lark-minutes

Search, download, and edit Feishu Minutes recordings and AI artifacts via lark-cli.

3|2|Updated Jul 21, 2026
One-click install
npx skills add https://github.com/zlsbksdxl/codex-lark --skill lark-minutes-zlsbksdxl
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: lark-minutes
Source: https://github.com/zlsbksdxl/codex-lark/tree/main/plugins/feishu2codex/skills/lark-minutes
Command: npx skills add https://github.com/zlsbksdxl/codex-lark --skill lark-minutes-zlsbksdxl

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Managing Feishu Minutes (妙记) manually—searching recordings, downloading media, extracting transcripts, and editing AI-generated summaries or todos—requires navigating multiple web interfaces. This Skill lets an AI agent perform all of these operations through structured lark-cli commands. ## Core Features & Use Cases - Search & Inspect Minutes: Find minutes by keyword, owner, participant, or time range, and retrieve titles, durations, and linked meeting notes. - Media & Artifact Handling: Download audio/video recordings, upload local media to generate new minutes, and fetch AI artifacts like summaries, todos, chapters, keywords, and transcripts. - Edit Minutes Content: Rename minutes, replace the AI summary, manage AI todos, replace transcript speakers, and batch-replace transcript keywords. - Use Case: A user says "turn this mp3 recording into a transcript with a summary." The agent uploads the file to Drive, converts it to a Minute, waits for processing, and returns the transcript and AI summary. ## Quick Start Ask the agent to search your recent Feishu minutes and show the summary and transcript of the most relevant one.

Frequently Asked Questions about lark-minutes

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I search Feishu Minutes by keyword or date?▼

Use lark-cli minutes +search with --query for keywords, --start/--end for time ranges, and --owner-ids or --participant-ids for people filters. At least one filter condition is required, and the time range is limited to one month per query.

How to convert a local audio or video file into a Feishu Minute transcript?▼

First upload the file with lark-cli drive +upload to get a file_token, then run minutes +upload --file-token to create the Minute, and finally call minutes +detail --wait-ready --transcript to fetch the transcript. Do not use local tools like ffmpeg or whisper.

What audio and video formats does Feishu Minutes upload support?▼

Supported audio formats include wav, mp3, m4a, aac, ogg, wma, and amr; video formats include avi, wmv, mov, mp4, m4v, mpeg, ogg, and flv. Files must be under 6 hours long and 6 GB in size.

Why does minutes download fail with error 2091003?▼

Error 2091003 means the Minute has not finished transcription yet, so its media file is not available for download. Wait for processing to complete and retry the download.

Can I edit AI todos inside a Feishu Minute instead of Feishu Tasks?▼

Yes, use minutes +todo with add, update, or delete operations to manage AI todos within a specific Minute. This is separate from Feishu Tasks; never route Minute todos through the task commands.

How do I fix a wrongly identified speaker in a Minutes transcript?▼

First query the speaker list via lark-cli api GET on the transcript speakerlist endpoint to get the speaker_id, resolve the new speaker's open_id via contacts, then run minutes +speaker-replace with --from-speaker-id and --to-user-id.