Apps & Interfaces · 07.02
The Datastore & the Data Loop
Filter/sort/limit/cursor queries with optional strict schemas, live DATA_CHANGED updates, and terminal run outcomes feeding back into records.
Closing the loop between what a run did and what the app shows
The datastore exposes ordinary filter, sort, limit, and cursor queries over an App's collections, with an optional strict schema per collection. What makes it a loop rather than just a table is that mutations broadcast a DATA_CHANGED event over the workspace realtime room — a grid open in the dashboard updates live the instant an agent writes a row, no manual refresh — and a run's terminal outcome feeds back into records and subjects automatically, so the automation's result and the product's displayed state are never two separate things that can silently drift apart.
agentis.data.define_collection · insert · upsert · update · delete · query · promote_memoryContinue
An App bundles typed collections, agent-authored surfaces, and orchestration — the product primitive a standalone workflow auto-wraps into.
An App is created with an auto-assembled cast of specialist agents — reused from the workspace talent pool where a fit exists, materialized where it doesn't.