Enterprise API Reference

HuMem.cloud exposes an extremely robust, RESTful API layer perfectly suited for integration with Latenode workers, multi-agent frameworks (CrewAI, AutoGen), and backend systems (Node.js, Python).

Authentication

All endpoints require a Bearer token. Generate your HUMEM_API_KEY from the dashboard and pass it in the Authorization header.

Authorization: Bearer YOUR_HUMEM_API_KEY

Core Endpoints

POST /v1/memory/ingest

Stores a single fact or experience into the temporal knowledge graph.

Payload:

{
  "user_id": "agent-orchestrator-1",
  "role": "user",
  "content": "The CEO prefers dark mode dashboards.",
  "scope": "tenant-local", /* Optional: 'tenant-local' or 'global-shared' */
  "modality": "text" /* Optional: 'text', 'image', 'audio', 'video' */
}

POST /v1/memory/ingest/batch

Enterprise batch ingestion. Ideal for Latenode workers scraping web context and dumping it instantly into memory. Uses a synchronous SQLite transaction for maximum throughput.

Payload:

{
  "user_id": "scraper-bot",
  "memories": [
    { "role": "system", "content": "Context block 1..." },
    { "role": "system", "content": "Context block 2..." }
  ]
}

POST /v1/memory/query

Perform a hybrid search (semantic vector + temporal SQLite relational) across your specific cognitive scopes.

Payload:

{
  "user_id": "agent-orchestrator-1",
  "query_string": "What does the CEO prefer for the UI?",
  "category": "left", /* Optional: filter by cognitive layer */
  "scope": "tenant-local"
}

Response:

{
  "results": [
    {
      "content": "The CEO prefers dark mode dashboards.",
      "score": 0.94,
      "timestamp": "2026-07-03T12:00:00Z"
    }
  ]
}

Advanced Integrations

POST /v1/orca/*

Direct proxy to the OrcaOS intelligence layer for heavy multimodal processing, pre-emptive summarization, and cognitive load resolution.

POST /v1/linea/*

Direct proxy to the Linea-1 orchestration layer. Pass workloads directly to the dynamic workflow engine to apply graph fixes or run complex dispatch routines.