arrowspace-mcp 0.3.0

MCP server for ArrowSpace spectral vector search

ArrowSpace MCP

MCP (Model Context Protocol) server for ArrowSpace spectral vector search.

Exposes ArrowSpace's graph-Laplacian-based spectral search as MCP tools for AI agents.

Quick start

uv run arrowspace-mcp

Client configuration

Claude Desktop

{
  "mcpServers": {
    "arrowspace": {
      "command": "uvx",
      "args": ["arrowspace-mcp"]
    }
  }
}

Cursor

Add to Cursor MCP configuration:

{
  "mcpServers": {
    "arrowspace": {
      "command": "uvx",
      "args": ["arrowspace-mcp"]
    }
  }
}

OpenCode

{
  "mcpServers": {
    "arrowspace": {
      "command": "uvx",
      "args": ["arrowspace-mcp"]
    }
  }
}

Transports

Transport	Command
stdio (default)	uv run arrowspace-mcp
SSE	uv run arrowspace-mcp --transport sse --port 8765

CLI flags

Flag	Default	Description
--transport	stdio	Transport protocol (stdio or sse)
--host	127.0.0.1	Host to bind (SSE only)
--port	8765	Port to bind (SSE only)
--allowed-paths	.	Colon-separated allowed directories for file_path
--ttl	3600	Index TTL in seconds
--max-indexes	100	Maximum concurrent indexes

Tools

Tool	Description
build_index	Build a spectral index from vectors (inline or file)
search	Query with λτ spectral gating
lambdas	Per-item λτ scores with optional summary stats
lambdas_sorted	Sorted λτ scores (least to most coherent)
spectral_analysis	Laplacian eigenvalue spectrum, components, or clusters
delete_index	Remove an index from memory

SSE deployment

uv run arrowspace-mcp --transport sse --host 0.0.0.0 --port 8765

Connect via http://<host>:8765/sse with messages at http://<host>:8765/messages/.

Development

git clone https://github.com/Genefold/arrowspace-mcp
cd arrowspace-mcp
uv pip install -e .
uv run arrowspace-mcp

With optional deps for spectral analysis

uv pip install -e ".[spectral]"

License

Apache 2.0