API Reference
Use the OpenGateway API base URL and Bearer authentication. Browse data-plane endpoints, check current creator-prefixed model IDs, and find the error reference.
Base URL#
Use the OpenGateway data-plane base URL for API requests:
https://apis.opengateway.ai/v1
Authentication#
Authenticated endpoints use an OpenGateway API key as a Bearer token:
Authorization: Bearer YOUR_OPENGATEWAY_API_KEY
GET /v1/models is public and does not require a key. /v1/messages also accepts x-api-key when no Authorization header is sent. See Authentication for details and API keys to create a key.
Model IDs and catalog#
Model IDs use the creator/model pattern. Creator namespaces include openai/, anthropic/, google/, x-ai/, moonshotai/, deepseek/, z-ai/, qwen/, minimax/, sionic-ai/, and typesafe/. Current examples include openai/gpt-5-mini, anthropic/claude-sonnet-4-6, and google/gemini-2.5-flash.
Use GET /v1/models to inspect current model IDs and their supported endpoints. A key restricted to an allowlist returns only models allowed for that key. The model catalog at /models includes current model details and pricing.
Endpoints#
Use the endpoint listed for a model in the catalog.
| Method | Path | Reference |
|---|---|---|
POST | /v1/chat/completions | Chat Completions |
POST | /v1/responses | Responses |
POST | /v1/messages | Messages |
POST | /v1/embeddings | Embeddings |
POST | /v1/images/generations | Images |
POST | /v1/images/edits | Images |
POST | /v1/decisions | Decisions |
GET | /v1/models | Models |
/v1/messages uses the Anthropic Messages API format and is available only for models that list messages in their supported endpoints.
Errors#
For error response formats and common error codes, see the Errors guide.