This overview answers the three questions most teams have before integrating: where to start, how to reason about limits, and where to debug failures.
Navigation Overview
Choose the capability line first
Start Here Decide whether your work is text, video, or image first, then go directly to the matching protocol or capability page. Open page What you will find:- Chat / agents -> text
- Media generation -> video or image
- Unified platform -> compare protocols first
Understand rate limits
Limits Treat request windows, concurrency, and output size as engineering constraints, not cleanup work for later. Open page What you will find:- Request windows
- Concurrency ceilings
- Over-limit handling
Trace request failures
Debug Use the error-code guide to quickly separate auth issues, parameter mistakes, rate limits, and server-side failures. Open page What you will find:- Auth failures
- Validation errors
- Server exceptions
Suggested reading order
- Confirm whether your use case belongs to text, video, or image capabilities.
- Open the specific protocol or capability page you actually need and integrate against that request shape.
- Use the rate-limit and error-code pages to finish the engineering constraints and debugging workflow.
Integration guidance
- If you already know the provider, go directly to that provider’s text, image, or video section.
- If you are still shaping the workflow, start from the usage examples and then open the matching endpoint page.
- Normalize provider-specific differences on the server if you are building a unified platform layer.