AGKit vs Your own backend
Choose AGKit when you want a managed backbone for application AI: model routes, profiles, credentials, budgets, and usage controls connected to your app’s tools.
A custom backend gives you control over model access, routing, identity, and operations. Your team builds and maintains each part of that service.
 AGKit | Your own backend | |
|---|---|---|
| Application stack | Swift, Android, React Native, and Web; one Cloud backbone | Your chosen client and server stack |
| Application actions | Agent runtime, tool contracts, permission hooks, and cancellation | Your own orchestration and application integration |
| Configuration | Cloud-managed model tiers, credentials, and agent profiles | Your model-routing implementation |
| Production controls | End-user identity, project/user budgets, and configured attestation | Your identity, budget, and operational controls |
| Application responsibility | Tool implementation, state validation, approvals, and actual undo | Your deployment, monitoring, and maintenance |
Choose your platform and access path for SDK availability and setup. Guard and confirmation support depends on the SDK execution path you integrate.
Where a custom backend fits
You need infrastructure, deployment, or policy behavior specific to your product, and have a team to own it. You control the architecture and its operational tradeoffs.
Where AGKit fits
You want model routing, profiles, credentials, budgets, and usage controls as a managed service connected to your application runtime.
What your team implements
With AGKit, focus on your product’s tools, state, and user experience. With a custom backend, also plan for provider integration, authentication, streaming, metering, monitoring, and incident response.