Getting Started
Next Steps
After add and query work, the next real choice is which route family you need next: deeper retrieval, project and key management, external data ingest, or enterprise identity.
If you need more than add and query
| Route | Why to read it next |
|---|---|
/memory/summarize | Summarize a known set of memory ids. |
/memory/compose | Build a prompt payload from retrieved memories and a token budget. |
/memory/search/routed | Use the routed-query pipeline when you want explainability-rich search behaviour. |
/memory/export | Export memories for archival or debugging. |
/memory/user/purge | Delete all memory data for the authenticated user. |
If you need richer read models
| Endpoint | Verified read model |
|---|---|
/memory/graph | Graph response. |
/memory/clusters | Clusters response. |
/memory/decisions | Decision panel response. |
/memory/intelligence | Intelligence response. |
If you need organization and project controls
- Project CRUD is mounted under
/org/projects. - API-key management is mounted under
/org/api-keys. - Continue sending
X-Project-IDwhenever the key is not pinned to one project.
If you need external data ingest
The verified integration providers are GitHub, Notion, Google Drive, and OneDrive.
- Integration routes are mounted under
/api/v1/integrations. - The web crawler is mounted under
/api/v1through the crawler router. - Org-level webhook management lives under
/org/webhooks.
If you need enterprise identity
- SSO routes exist for both SAML and OIDC login flows.
- SCIM provisioning is mounted at
/scim/v2/*. - OAuth browser authorization starts at
/oauth/authorize.
If you are operating the system
- Billing routes are mounted under
/org/billing.