hm_get_nodes

Use hm_recall to find memories and hm_get_nodes to read them in full. It returns each requested node with its complete description, data, relationships and hyperedges.

Parameters

Parameter	Type	Default	Description
`keys`	string[]	required	1–50 exact node keys.
`include_relationships`	boolean	`true`	Include the node’s binary edges and hyperedges.
`relationship_detail`	`"summary"` \| `"full"`	`"summary"`	How much of each relationship to return.
`participant_limit`	integer	`20`	How many participants to list per hyperedge (0–100).
`include_duplicates`	boolean	`true`	Include records that were merged into this node.
`include_enrichment_history`	boolean	`false`	Include the history of background enrichment passes.
`include_session_relationships`	boolean	`false`	Include the server’s chat-session groupings.
`organization_id`	string	—	Read an organization graph you belong to instead of your own. The result carries `read_target`.

Response

Field	Description
`nodes`	An object keyed by node key. Each entry has `node` (key, type, description, data, enrichment status, `write_version`), `binary_edges` (each with its `id`, `from_key`, `to_key`, `relationship`, `write_version`), `hyperedges`, and `duplicate_records`.
`missing`	Keys that don’t exist.
`count`	Number of nodes returned.

A node with more relationships than one read covers returns the ones that fit, with relationships_truncated: true.

The id of each edge and its write_version are what hm_update and hm_forget take as edge_id and edge_version.

Example

{
  "keys": ["decision_jwt_auth", "project_public_api"],
  "relationship_detail": "full"
}

Equivalents

curl -X POST https://api.hypermemory.io/api/v1/memory/nodes \
  -H "Authorization: Bearer $HYPERMEMORY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"keys":["decision_jwt_auth"]}'