Skip to main content
Graph Memory enriches Recallio with a knowledge graph that links people, locations, listings, and processes. Use these pages to plan how structured relationships support richer answers and workflows.

Core capabilities

  • Ingest graph facts with POST /api/GraphMemory/addv2, automatically extracting entities and relationships from structured JSON or text.
  • Query connected context using POST /api/GraphMemory/searchv2 to surface the most relevant paths for an assistant prompt.
  • Inspect the graph via GET /api/GraphMemory/relationships to understand how nodes connect for a user or project.
  • Reset scope with DELETE /api/GraphMemory/delete-all (Manager role) when you need to remove stale relationships.
All Graph Memory endpoints share the same bearer token you create in Dashboard → API Keys. Use a key with the Manager role for delete operations.

Endpoint quick reference

Use caseEndpointMethodKey inputs
Add entities and relationships/api/GraphMemory/addv2POSTBody: data (JSON or text), optional user_id, project_id
Search graph for relevant paths/api/GraphMemory/searchv2POSTBody: query, optional user_id, project_id, limit, threshold
List relationships for inspection/api/GraphMemory/relationshipsGETQuery params: userId, optional projectId, limit
Delete all relationships for a scope/api/GraphMemory/delete-allDELETEQuery params: userId, optional projectId (Manager key required)

Before you start

  1. Generate an API key and choose the appropriate role. Contributor covers add/search; Manager is required for delete.
  2. Decide on scoping. Populate user_id when graph data belongs to a specific resident or agent, and project_id to isolate relationships per environment.
  3. Structure your payloads. Consolidate related attributes into a JSON object under the data field so Recallio can detect entities and relations automatically.

Where to go next

Review full schemas in the Recallio Swagger reference.