Skip to main content
A database is a physically isolated store for context. In most integrations you create a database once, wait for provisioning, then ingest and query inside it.

Endpoint references

SDK methods are listed by their Python names. The TypeScript SDK uses the same names in camelCase, such as databases.deleteCollection. The asynchronous calls return before the work finishes. Each endpoint page explains how to confirm completion.

Typical call sequence

  1. Create the database: POST /databases.
  2. Wait for provisioning: poll GET /databases/status until infra.ready_for_ingestion is true.
  3. Ingest context and wait for indexing: see the Context overview.
  4. Query: see the Query overview.
  5. Inspect usage: GET /databases/stats returns row counts for knowledge and memories.

Key concepts