feishu-doc-cli

CLI tool to read Feishu Open Platform documentation as Markdown.

Feishu developer docs live behind a SPA that AI coding agents cannot read directly. This tool exposes the docs through a simple CLI, outputting raw Markdown that agents (and humans) can consume.

Install

npm install -g feishu-doc-cli

Requires Node.js >= 18.

As an AI agent skill

npx skills add m1heng/feishu-doc-cli

Once installed, agents like Claude Code will automatically use feishu-doc when you ask about Feishu/Lark APIs.

Usage

Read a document

feishu-doc read /home/intro

Output:

# 开放平台概述

> Path: /home/intro
> Updated: 2025-11-28

---

(original Markdown content)

Browse the document tree

feishu-doc tree --depth 2

├── 📁 文档首页
│   └── 📄 首页  → /home/index
├── 📁 开发指南
│   ├── 📁 平台简介
│   ├── 📁 开发流程
│   ...

Filter to a subtree:

feishu-doc tree "/uAjLw4CM/uYjL24iN/platform-overveiw" --depth 3

Search documents by title

feishu-doc search "消息"

Found 109 document(s):

  发送消息
    feishu-doc read "/uAjLw4CM/ukTMukTMukTM/reference/im-v1/message/create"

Options

| Option | Description | Default | |--------|-------------|---------| | --lang zh\|en | Language | zh | | --depth <n> | Tree display depth | unlimited |

Following links

Documents contain links in various formats. All of them work directly with feishu-doc read — just copy the href value from any link in the content:

# /ssl:ttdoc/ links (most common in content)
feishu-doc read "/ssl:ttdoc/ukTMukTMukTM/uMTNz4yM1MjLzUzM"

# Full Feishu URLs
feishu-doc read "https://open.feishu.cn/document/client-docs/intro"

# Full Lark URLs
feishu-doc read "https://open.larkoffice.com/document/client-docs/bot-v3/bot-overview"

# /document/ paths
feishu-doc read "/document/client-docs/intro"

# API fullPath (as shown in tree output)
feishu-doc read "/home/intro"

The CLI normalizes all formats internally, including anchors (#section) and query strings. No need to manually convert.

How it works

Feishu Open Platform exposes two public APIs (no authentication required):

| API | Purpose | |-----|---------| | GET /api/tools/docment/directory_list | Full document tree (~925 KB JSON) | | GET /document_portal/v1/document/get_detail?fullPath=<path> | Document content in Markdown |

This CLI is a thin wrapper around these two endpoints.

License

MIT