About MCP Server
The @moss-tools/mcp-server package exposes Moss as a Model Context Protocol server, giving AI coding assistants like Claude Desktop, Cursor, and VS Code direct access to your knowledge base. Run it with npx - no installation required. The server provides tools for querying, indexing, and managing documents.
Why Use Moss with MCP Server
01
Zero install — run with npx
Run with `npx @moss-tools/mcp-server` — works immediately with Claude Desktop, Cursor, and VS Code
02
Full Moss API exposed as MCP tools
`query`, `create_index`, `load_index`, `add_docs`, `delete_docs`, and more
03
Conversational knowledge management
AI assistants can search your knowledge base, create indexes, and manage documents conversationally
04
Sub-10ms search latency
AI assistants get context without noticeable delay
05
Simple JSON configuration
Add to `claude_desktop_config.json` or `.cursor/mcp.json`
Quick Start
{
"mcpServers": {
"moss": {
"command": "npx",
"args": ["-y", "@moss-tools/mcp-server"],
"env": {
"MOSS_PROJECT_ID": "your-project-id",
"MOSS_PROJECT_KEY": "your-project-key"
}
}
}
}Get Started in 3 Steps
01
Add MCP server config
Add the Moss MCP server entry to your AI assistant's config file (claude_desktop_config.json for Claude Desktop, .cursor/mcp.json for Cursor).
02
Set environment variables
Provide your MOSS_PROJECT_ID and MOSS_PROJECT_KEY in the env section of the MCP config. The server authenticates with these credentials.
03
Start using tools
Restart your AI assistant. It can now query your Moss indexes, create new indexes, add documents, and manage your knowledge base through natural language.
Frequently asked questions
Ready to ship faster AI products?
Moss gives you production-ready semantic retrieval without infrastructure complexity.