API Reference

NexusModels provides a unified RESTful API for accessing multiple AI models.

The API is compatible with the OpenAI API format and supports text, embeddings, audio, images, video, and other model capabilities.

API availability

Supported endpoints and parameters vary by model. Use the Models API to retrieve the models available to your API key.

AI Model APIs

API	Common Endpoint	Description
Models	GET /v1/models	List models available to the current API key
Chat	POST /v1/chat/completions	Create multi-turn chat completions with standard or streaming output
Responses	POST /responses	Create a model response using the OpenAI Responses format
Completions	POST /completions	Create a legacy text completion
Embeddings	POST /embeddings	Convert text into vector representations
Rerank	POST /rerank	Rerank documents by relevance
Moderations	POST /moderations	Analyze input for potentially unsafe content
Realtime	Realtime WebSocket API	Real-time text and audio interaction
Audio	/audio/speech, /audio/transcriptions, /audio/translations	Text-to-speech, transcription, and translation
Images	/images/generations, /images/edits	Image generation and editing
Videos	Model-specific video endpoints	AI video generation

Endpoint availability depends on:

• Models available to the current API key
• Input and output types supported by the selected model
• Whether the model is currently enabled
• Upstream provider compatibility

Conventions

API Base URL

OpenAI SDKs and most API clients should use:

https://api.nexusmodels.cn/v1

Full endpoint example:

https://api.nexusmodels.cn/v1/chat/completions

Authentication

All model requests must include a NexusModels API key:

Authorization: Bearer YOUR_NEXUSMODELS_API_KEY

Example:

curl https://api.nexusmodels.cn/v1/models \
  -H "Authorization: Bearer YOUR_NEXUSMODELS_API_KEY"

Keep your API key secure. Do not:

• Commit it to a public repository
• Include it in browser-side code
• Print it in public logs
• Share it with unauthorized users

Request Format

GET requests normally use query string parameters:

GET /v1/models

POST requests normally use a JSON body:

Content-Type: application/json

Example:

{
  "model": "gpt-5.4",
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ]
}

File upload endpoints may use:

Content-Type: multipart/form-data

Response Format

Most endpoints return JSON data.

Common response fields include:

Field	Type	Description
id	string	Unique request or response identifier
object	string	Object type
model	string	Model used for the request
created	integer	Response creation timestamp
choices	array	Generated results
usage	object	Token usage information

Example chat response:

{
  "id": "chatcmpl-example",
  "object": "chat.completion",
  "created": 1780000000,
  "model": "gpt-5.4",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello. How can I help you?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 10,
    "completion_tokens": 15,
    "total_tokens": 25
  }
}

Streaming endpoints return Server-Sent Events:

Content-Type: text/event-stream

Audio and image endpoints may return file URLs or binary content depending on the selected model and endpoint.

Quick Requests

List Models

curl https://api.nexusmodels.cn/v1/models \
  -H "Authorization: Bearer YOUR_NEXUSMODELS_API_KEY"

Create a Chat Completion

curl https://api.nexusmodels.cn/v1/chat/completions \
  -H "Authorization: Bearer YOUR_NEXUSMODELS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "messages": [
      {
        "role": "user",
        "content": "Hello"
      }
    ]
  }'

Create a Streaming Request

curl https://api.nexusmodels.cn/v1/chat/completions \
  -H "Authorization: Bearer YOUR_NEXUSMODELS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "messages": [
      {
        "role": "user",
        "content": "Explain artificial intelligence."
      }
    ],
    "stream": true
  }'

Error Handling

API errors normally use the following format:

{
  "error": {
    "message": "Error description",
    "type": "error_type",
    "param": null,
    "code": "error_code"
  }
}

Common HTTP status codes:

HTTP Status	Error Type	Description
400	invalid_request_error	Invalid request parameters or format
401	authentication_error	Missing, invalid, or disabled API key
403	permission_error	Access to the requested model or endpoint is denied
404	not_found_error	Model or endpoint not found
429	rate_limit_error	Rate, usage, or account limit exceeded
500	api_error	Internal NexusModels error
502	upstream_error	Upstream model provider error
503	service_unavailable	Service temporarily unavailable

Clients should use the HTTP status code to determine whether a request should be retried. For 500, 502, and 503, use a limited exponential backoff strategy.

View the complete error handling guide