Real Memory AI
Clients
Create models, upload approved sources, understand evidence, manage access, and use RM AI without hidden execution claims.
Documentation you can verify
Know what runs, what is governed, and what is still planned.
The live documentation covers model creation, source ingestion, grounded chat, tenant roles, feature entitlements, approval semantics, billing, privacy, local validation, and production delivery.
Developers
Review scoped authentication, endpoint behavior, rate limits, safe outbound fetching, and tenant-aware data contracts.
Operators
Use local-first checks, immutable release planning, readiness gates, backup manifests, rollback procedures, and publish approvals.
Historical completion notes are not the runtime contract. Current capability catalogs, tests, readiness checks, and blocker records take precedence.