English | 中文

Lark Wiki Migration

A cross-AI-product migration tool for Feishu (Lark) wiki spaces. Give it a source wiki link and a target wiki link, and it migrates all docx documents from the source space to the target, preserving the tree structure.

Available through:

• CLI: lark-wiki-migrate <source> <target>

• Claude Code skill: triggered by natural language

• MCP Server: any MCP-compatible AI product (Codex, Workbuddy, etc.)

Prerequisites

1. Install lark-cli and log in:

   lark-cli auth login

2. The logged-in account must have both:

   • Read permission on the source wiki space

   • Create document permission on the target wiki space

Related MCP server: lark-mcp

Installation

git clone https://github.com/LiChunyu77/lark-wiki-migration.git
cd lark-wiki-migration
./install.sh

install.sh will:

• Symlink the lark-wiki-migrate command into ~/.local/bin

• Automatically install the Claude Code skill if ~/.claude/skills is detected

After installation, make sure ~/.local/bin is in your PATH:

export PATH="$HOME/.local/bin:$PATH"

CLI Usage

# Migrate
lark-wiki-migrate "https://my.feishu.cn/wiki/SOURCE" "https://my.feishu.cn/wiki/TARGET"

# Preview only, create nothing
lark-wiki-migrate "https://my.feishu.cn/wiki/SOURCE" "https://my.feishu.cn/wiki/TARGET" --dry-run

# Skip confirmation
lark-wiki-migrate "https://my.feishu.cn/wiki/SOURCE" "https://my.feishu.cn/wiki/TARGET" --yes

Claude Code Usage

Restart Claude Code after installation, then say:

Migrate https://my.feishu.cn/wiki/SOURCE to https://my.feishu.cn/wiki/TARGET

Claude will automatically invoke lark-wiki-migrate and return the result.

MCP Server Usage

Works with any MCP-compatible AI client.

1. Configure the MCP server:

   {
     "mcpServers": {
       "lark-wiki-migration": {
         "command": "python3",
         "args": ["/absolute/path/to/mcp_server.py"]
       }
     }
   }

2. The client will expose a migrate_feishu_wiki tool that accepts:

   • source: source wiki URL or token

   • target: target wiki URL or token

   • dry_run: optional, scan only without creating anything

How It Works

• lark-cli docs +fetch reads the source document XML

• lark-cli docs +create --doc-format xml creates native docx documents under the target wiki

• Recursively walks the source wiki tree, preserving the hierarchy

Output

Each migration creates a dedicated working directory under ~/lark-wiki-migrations/ by default:

~/lark-wiki-migrations/
└── wiki-migration-<slug>-<timestamp>/
    ├── manifest.json
    ├── source_tree.json
    └── migration_state.json

• source_tree.json: source wiki structure

• migration_state.json: resumable migration state

• manifest.json: metadata for this migration run

Environment Variables

export LARK_WIKI_MIGRATION_DIR=~/my-migrations   # Custom working directory base path

Notes

• Only docx documents are migrated; embedded spreadsheets, bitables, whiteboards, and similar objects are not fully migrated

• Images and attachments are usually preserved, but access-restricted resources may fail

• The tool only uses the local lark-cli login session; it does not collect, upload, or store any credentials

• The tool operates through official APIs and can only access content your account already has permission to access — it does not bypass any permission restrictions

• Compared with manually copying/exporting documents one by one, this tool supports batch migration, preserves the wiki hierarchy, and can resume interrupted migrations

• Make sure you have legitimate access to both the source and target wiki spaces

• Use of this tool must comply with the Feishu Open Platform terms of service and your organization's IP policies

License

MIT