Application types
Last updated
Last updated
The OXU Developer ortalp supports three application types: service application, AI application, and game application. You choose a type when creating an app, and you can adjust resources and permissions later.
Default resource templates in manage app’s resources (client identity, app capabilities and security, ecosystem connection)
Suggested scopes/permissions and security defaults
Type-specific fields in app info and Store categorization
Review guidelines applicable to your listing
Best for business and productivity integrations, backend services, workflow automation, and connectors.
Typical capabilities:
Expose REST/Graph APIs via a Resource Server
Support webhooks/events for sync or notifications
Authenticate users via SPA or Regular Web OAuth clients; optional M2M Integrate Client for server-to-server
Notes:
Start with least-privilege scopes; request only what you need
Ensure idempotency and clear error codes for API endpoints
Service: Resource Server (business APIs), SPA/Regular OAuth client, optional Integrate (M2M), webhooks
AI: Resource Server (inference endpoints), Regular/SPA OAuth client, Integrate (M2M) for jobs, webhook callbacks
Game: SPA/Native OAuth client, optional backend Resource Server, commerce-related webhooks
All endpoints must return stable status codes and minimal contract shapes
Scopes must be enforced server-side; deny by default
Handle retries and timeouts gracefully; include request IDs in logs
Keep Sandbox data isolated from production
Best for apps that provide AI/ML powered features like generation, classification, embeddings, or retrieval.
Typical capabilities:
Notes:
Best for gaming and entertainment experiences that integrate OXU identity and commerce.
Typical capabilities:
Notes:
Define resources (APIs, clients, webhooks) under Manage App’s Resources
Provide App Info, assets, and marketplace details
Optionally set pricing: see Manage Pricing in the Developer Portal
Create an app and select a type:
Configure authentication and permissions: ,
Test end-to-end in Sandbox:
Submit for review:
Model-backed inference endpoints; jobs may be synchronous or asynchronous with callbacks
User-facing OAuth client (SPA/Regular) and optional M2M client for batch jobs
Usage- and privacy-aware logging; avoid sensitive data in logs
Document your model/provider and expected inputs/outputs
Consider rate limits and retries for long-running operations
Player identity via OAuth (SPA/Native), optional backend for authoritative state
Optional webhooks for purchases, achievements, or telemetry
Pricing models for in-app purchases or subscriptions
Follow content policies and age ratings where applicable
Keep client secrets on server-side only; never in client binaries