neruva-mcp 0.3.0

Model Context Protocol server for Neruva — vector memory + HD-native substrate (KGs, analogy, causal do-operator, EFE planning).

neruva-mcp

Model Context Protocol server for Neruva — vector memory + HD-native substrate. Same 8 tools as the TypeScript package, for Python-only MCP hosts (e.g. Codex, custom agents).

Install

pip install neruva-mcp
export NERUVA_API_KEY=...        # from https://app.neruva.io

Or zero-install via uvx:

uvx neruva-mcp

Use

Add to your MCP host's config. Example for Claude Code:

claude mcp add neruva -- uvx neruva-mcp -e NERUVA_API_KEY=YOUR_KEY

Or run directly:

NERUVA_API_KEY=YOUR_KEY neruva-mcp

Tools

Tool	What it does
memory_upsert	Insert/update vectors in a namespace (Pinecone-compatible)
memory_query	Top-K cosine search
memory_delete	Delete by id
hd_kg_add_fact	Add (subject, relation, object) to a KG
hd_kg_query	Query KG for object of (subject, relation)
hd_analogy	Four-term analogy in HD feature space
hd_causal_query	Observational / interventional query on stored SCM
hd_plan	EFE active-inference planner

Env

Variable	Default
NERUVA_API_KEY	required
NERUVA_URL	https://api.neruva.io

License

MIT — Clouthier Simulation Labs.