hm_store

Create a new node in the hypergraph. After storing, the server triggers background enrichment — an LLM pass that expands the description, extracts structured data, creates new entities and edges, and updates the user_profile singleton.

Parameters

Parameter	Type	Default	Description
key	str	required	Unique identifier for the node. Use {type}_{name} format.
description	str	required	What this memory represents. Searched by BM25 and vector similarity.
data	dict | None	None	Structured metadata (JSON key-value pairs).
assets	list[str] | None	None	Optional asset references attached to the node
node_type	str | None	None	Type classification (e.g. person, decision, concept).
relationships	list[dict] | None	None	Edges/hyperedges to create alongside the node (see MCP schema / examples).

Enrichment

After a successful store, the server runs a background enrichment pipeline that:

1. Expands the description with additional context
2. Extracts structured data fields
3. Auto-detects entities mentioned in the description and creates them as new nodes
4. Creates relationship edges between the new node and detected entities
5. Updates the user_profile singleton with relevant information

Response

Returns the created node key and confirmation. The enrichment pipeline runs asynchronously — enriched data appears on subsequent recall.

Example

{
  "key": "decision_jwt_auth",
  "description": "We chose JWT over session cookies for stateless scaling",
  "node_type": "decision",
  "data": {"status": "approved", "date": "2026-04-01"}
}

CLI equivalent

hm store decision_jwt_auth "We chose JWT over session cookies" --type decision

REST endpoint

POST /api/v1/memory/store