V6.4

CLAIV Memory docs

CLAIV Memory is a context engine for AI applications. Ingest conversation events; recall pre-synthesized context; let your LLM generate the reply.

Your tenant is inferred from your API key — you never send a tenant_id in payloads.

npm install @claiv/memory

pip install claiv-memory

Quickstart

Install an SDK, ingest your first event, recall context, and forget data in under 5 minutes.

Concepts

Memory lifecycle, block types, conflict detection, scoping, tiered memory, and role-aware extraction.

API Reference

Complete V6.4 docs for /ingest, /recall, /forget, /documents, /collections, usage, and health endpoints.

Usage & Limits

Monitor ingests and recalls, understand quotas, overage behaviour, and rate limits.

Security

API key lifecycle, tenant isolation, input validation, encryption, and data protection.

Deployment

Deploy the CLAIV Memory backend: API server, worker, Postgres, and Redis.

Troubleshooting

Common issues, HTTP error codes, debugging tips, and support escalation.

How it works in 30 seconds

1Ingest — call POST /v6/ingest with a message. Returns immediately; facts extract asynchronously in 1–5 s.
2Recall — before your LLM call, ask POST /v6/recall. Get back llm_context.text — inject it directly into your system prompt.
3Forget — call POST /v6/forget for GDPR right-to-erasure. Returns an auditable receipt of everything deleted.