Lanhu MCP Server

Lanhu MCP Server is a team collaboration MCP server designed for the AI coding era.

Core Capabilities:

• Requirement Document Analysis: Automatically parse Axure prototypes with three analysis modes (dev/test/quick explore), four-stage workflow for zero omissions, >95% accuracy
• UI Design Support: Batch download design files, auto-extract slices, intelligent semantic naming
• Team Collaboration Message Board: Unified knowledge base for all AI assistants to share context, with task collaboration, @Feishu notifications, collaborator tracking
• Performance Optimization: Version-based smart caching, incremental updates, concurrent processing

MCP Tools:

• lanhu_resolve_invite_link: Parse invite links
• lanhu_get_pages / lanhu_get_ai_analyze_page_result: Get and analyze prototype pages
• lanhu_get_designs / lanhu_get_ai_analyze_design_result: Get and analyze UI designs
• lanhu_get_design_slices: Get slice information
• lanhu_say series: Message post/view/edit/delete
• lanhu_get_members: View collaborators

Message Types:

Deployment:

• Docker (Recommended): git clone → setup-env.sh → docker-compose up -d
• Source: easy-install.sh → python lanhu_mcp_server.py
• Default server address: http://localhost:8000/mcp

Required Configuration:

• LANHU_COOKIE: Obtain from Lanhu web version
• Optional: FEISHU_WEBHOOK_URL (Feishu notifications)

Compatible AI Clients: Cursor, Windsurf, Claude Code, Tongyi Lingma, Cline

Notes:

• Must use AI models with vision capabilities (Claude, GPT, Gemini, Kimi, Qwen, DeepSeek, etc.)
• Third-party open-source project (MIT License), not officially affiliated with Lanhu
• Data processed and cached locally, not transmitted to third-party servers