Apps Overview
Apps is the section of your project where you generate full-stack applications from a natural language prompt — a capability qRaptor calls Vibe Code. It lives in the project sidebar under Apps, and its landing page heading reads Create Apps.
What you get is a complete transactional business application, not a prototype or a chat wrapper: a relational PostgreSQL data model, a REST API over it, authentication, roles and page-level permissions, business rules, workflows and a full UI — the kind of system a team runs its operations on. AI agents are optional components of that application, not a prerequisite for it; an app with no agents at all is a normal thing to build here.
What you can build
Every app starts the same way: you describe what you want, and qRaptor generates it. The platform selector on the Apps landing page offers three types:
| Type | What it produces | Notes |
|---|---|---|
| Web | A full-stack web app (Postgres-backed) | Default selection |
| Mobile | A full-stack mobile app (Postgres-backed) | Marked Preview — mobile app generation is still in preview |
| Copilot | An embeddable AI chat widget bound to an agent or MAS system | Uses a guided 2-step wizard instead of a free-text prompt |
Web and Mobile share the same prompt composer and generation pipeline — see Vibe Coding. Copilot uses its own guided creation wizard instead — see Copilot.
Web and Mobile apps use a postgresql database by default. Copilot apps don’t provision a database — they connect to an existing agent or MAS system as their “brain.”
The app lifecycle
Every app follows the same lifecycle, regardless of type:
- Create — Pick a type (Web, Mobile, or Copilot), describe it, and generate.
- Generate — qRaptor turns your prompt into requirements, an app graph, and working code — fixing build issues automatically along the way.
- Review & iterate — Inspect the Requirements, Graph, Data Plan, API Doc, and Roles for your app. Chat with qRaptor to request changes, or embed an agent into your app.
- Preview — View your app running live, and sign in with test credentials if it’s authenticated.
- Secure — Check where authentication is configured for your app.
- Publish — Deploy your app to production with one click.
- Monitor — Check deployment status, view logs, and manage the running app.
Where each step lives
If you’ve used the Agentic Workspace before, the workspace concept here is similar — an Apps workspace is the full-screen builder for a single generated app, distinct from your qRaptor project.
| Concept | Where it lives |
|---|---|
| Create a new app | Apps landing page (/app-workspace) |
| The generation pipeline | The app workspace, opened after you generate |
| Code, Preview, App Details | Top-level tabs inside the workspace |
| Connect an agent to your app | Chat panel → + → Add agent integration |
| Connect a third-party service, such as payments | Apps → Connections |
| Enable authentication for an app | Auth chip, at creation |
| SSO, email server, MFA, login branding | Project sidebar → Security → Auth Settings |
| Publish | Publish button in the workspace toolbar |
| Deployment status | Project sidebar → Deployments |
What’s next
- Create an app → — Choose a type, write a prompt, and generate your first app
- The generation flow → — What happens between your prompt and a working app
- Preview your app → — View it live and sign in with test credentials
- Secure your app → — Where authentication is configured for your app
- Agent integrations → — Connect an agent as a chat, an action people trigger, or a call from your app’s server
- Publish to production → — Deploy your app
- Connections → — Connect your app to services such as Razorpay