Overview

The Memory Reconstruction API manages long-running or synchronous jobs that reconstruct an agent’s memory from session events, memory stores, or uploaded source material. Jobs are organization-scoped and require organization admin access.

Base URL

All endpoints require a valid Bearer token in the Authorization header and an active organization. See Authentication for details.

Endpoints

Create job

Request body:
Allowed source_type values: session, memory_store, file. Response:

List jobs

Response:

Run job

The synchronous runner produces a memory outline from the configured topics and source metadata. Future releases will support asynchronous worker execution with progress streaming. Response:

Delete job

Response:

Error codes