Independent, unofficial Cognee companion

Cognee Space

Plan persistent AI memory before agents go live: source ingestion, knowledge graph RAG, recall modes, MCP memory access, feedback loops, and self-hosted controls in one paid planner.

Pricing required

AI memory deployment planner preview

The planner is gated until a package is selected and Polar checkout is verified. The preview shows the required inputs without returning a full production plan.

{
  "status": "pricing_required",
  "next": "Choose a pricing package before using the planner.",
  "pricing": "/pricing/"
}

Page matrix

Built for high-intent AI memory searches

AI memory planning

Map what the agent should remember, where the data comes from, and which recall behavior belongs in the product.

Open AI memory

Knowledge graph RAG

Translate documents, entities, relationships, vector search, and graph reasoning into an implementation checklist.

Plan Graph RAG

MCP memory server

Decide when agents should access memory through stdio, SSE, HTTP, API mode, or a self-hosted Cognee service.

Review MCP setup