Interface
OpenAI-compatible
One entry point for production AI model calls
BatchIn Model API is OpenAI-compatible and keeps text, coding, reasoning, embeddings, image, audio, and video models under one catalog, routing layer, usage ledger, and billing path.
Status
Per-model state
Ledger
Usage / cost
Receipts
Verifiable
Core modalities for production calls
Model API is the unified entry; coding, embeddings, image, audio, and video are capabilities inside the API rather than separate top-level products.
- Chat, Responses, Completions, and Embeddings share the same auth and billing path.
- Image, audio, and video expose GA, evaluation, or commercial-review state from real model status.
- Request records flow into usage, traces, billing, and VaaS Ledger.
From developer trial to team production
Customers start with one base_url, then graduate through budgets, model policy, receipts, and capacity controls.
- Playground, Docs, Models, and Console use the same product truth.
- Pricing comes from runtime catalog or verified SKUs, not hardcoded frontend copy.
- Coding gateway is a Model API use case, not a separate first-level product.