Category 04 of 12
Agentic Apps
Introducing Agentic Apps — the App primitive that bundles typed data, agent-authored surfaces, a staffed cast, and orchestration into one durable, self-operating product.
Pages in this category
Not a chatbot, not a script, not a SaaS app with an AI feature bolted on. An Agentic App is durable software built from four real pillars — data, interface, orchestration, and its own Brain.
Every other app is frozen the day it ships. Give an Agentic App a Goal and it runs competing strategies toward it, measures which one actually wins, learns that in its own Brain, and evolves — promoting the winner and breeding the next generation. Under your control, and without lying about the results.
One end-to-end walk through what actually exists in the database and who touches it — from build_workflow to a staffed, data-backed, self-orchestrating product.
A real learning loop scoped to one App: outcomes become graded lessons, recurring lessons graduate into inherited skills — the same Brain formation path the rest of the platform uses, never a parallel system.
An App bundles typed collections, agent-authored surfaces, and orchestration — the product primitive a standalone workflow auto-wraps into.
Filter/sort/limit/cursor queries with optional strict schemas, live DATA_CHANGED updates, and terminal run outcomes feeding back into records.
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.
Live co-presence over the realtime bus, ephemeral by design, so two operators can tell they're looking at the same App at once.
Navigation, status, approvals, refresh, and an ops drawer that always resolves live active runs — how every App surface stays operable.
A real project-management backlog for agent work: status, priority, labels, recurrence, and assignment.
A .agentisapp v2 bundle carries the whole transitive App, an explicit revision target, and verified portable components.
Persisted, referenceable artifacts collected for delivery, with retention policy and a dedicated library surface.
A surface is a typed ViewNode tree — layout, data-bound views, interactions, and agent-native widgets an agent can render, patch, and lint safely.
How the surface generator classifies a collection's shape into an archetype — analytics, pipeline, CRM, roadmap, operations — and scaffolds a defensible default.
Run monitors, approval gates, agent feeds, orchestration panels, code viewers, media galleries, and maps — the widget vocabulary beyond tables and forms.
Semantic tones, palettes, and shell modes bound what an agent can style; the operability gate enforces that a rendered action is a wired action.
An App runs several workflows as a durable event-driven system: explicit entrypoints, verified-success rules, delivery evidence, retries, and recovery.