API overview

NexInfer currently publishes 12 text, multimodal, image, video, and speech models. The Anthropic Messages contract is retained for upcoming support and is not part of the current 12-model catalog.

Use case	Base URL	Primary endpoints
OpenAI SDK, Codex, Cline	https://api.nexinfer.com/v1	/responses, /chat/completions, /models
Image generation and edits	https://api.nexinfer.com/v1	/images/generations, /images/edits
Video generation	https://api.nexinfer.com/v1	/video/generations, task status, content download
Text to speech	https://api.nexinfer.com/v1	/audio/speech
Claude Code, Anthropic SDK (coming soon)	https://api.nexinfer.com	/v1/messages

Recommended validation order

1. Request /v1/models to verify the key and its model scope.
2. Send a minimal request through the client’s native endpoint.
3. Validate streaming, tool calls, and long context.
4. Add bounded timeouts and retries, then redact production logs.

NexInfer also accepts the historical /responses path. New configurations should use /v1/responses.

See the machine-readable OpenAPI JSON and live public models.json. Capability metadata comes from the model catalog; the authenticated /v1/models response remains authoritative for a particular key.