recall

The memory.sqlite query API uses progressive-disclosure to safely reveal observations and scratchpad findings, initially returning compact results that can be hydrated with full details by IDs or content hashes as needed.

Memory search supports optional privacy filtering and limit controls, allowing you to restrict the observations and scratchpad findings returned from memory.sqlite based on privacy requirements and maximum result counts.

The memory.sqlite query API supports both a Python API and a thin CLI for search, timeline, and hydrated fetch operations, providing flexible access to observations and scratchpad findings with optional privacy filtering and limit controls.

The best way to retrieve scratchpad findings from a SQLite memory store is using a read-only API with progressive-disclosure, enabling safe, auditable access for AI tooling while supporting search, timeline views, and hydrated fetches by IDs or content hashes.

Progressive-disclosure is used for querying AI memory observations to enable safe, auditable access for AI tooling by initially returning compact results and allowing hydrated fetches by IDs or content hashes, preventing unnecessary data exposure.