Live

API Reference — Gateway Endpoint Manual

Endpoint docs · error codes · request examples

API Overview Authentication Chat Completions

The Lurus API is fully compatible with the OpenAI API format, so you can call it directly with any OpenAI-compatible SDK or tool.

Why use the Lurus API

OpenAI compatible

All endpoint paths and semantics match OpenAI; swap in the base_url with zero SDK changes

Complete error codes

Every error carries code + message + a suggested action, making automated handling easy

Multiple auth methods

Bearer Token / PAT / JWT, covering everything from scripts to enterprise SSO

Model routing and retries

Automatically routes to upstream channels by model name and fails over to backups

Get started in three steps

1. Point the base URL at https://hub.lurus.cn/v1.

2. Include Authorization: Bearer sk-your-api-key in the request header (Get an API Key).

3. Send requests with any OpenAI SDK — no changes to your business code needed. See SDK support below.

Base URL

https://hub.lurus.cn/v1

Authentication

Every API request must include your API Key in the Header:

Authorization: Bearer sk-your-api-key

More than just API Keys

Beyond Bearer Token, OIDC / OAuth2, PAT, and JWT are also supported. See Authentication in detail.

Available endpoints

Get model list

GET/v1/models获取模型列表

GET/v1beta/modelsGemini 格式获取

OpenAI format (Chat)

POST/v1/chat/completions创建聊天对话

OpenAI format (Responses)

POST/v1/responses创建响应 (OpenAI Responses API)

POST/v1/responses/compact压缩对话 (OpenAI Responses API)

Image generation

POST/v1/images/generations生成图像(qwen-image)

POST/v1/images/edits编辑图像(qwen-image-edit)

Video generation

POST/v1/videos创建视频

GET/v1/videos/{task_id}获取视频任务状态

GET/v1/videos/{task_id}/content获取视频内容

POST/kling/v1/videos/text2videoKling 文生视频

GET/kling/v1/videos/text2video/{task_id}获取 Kling 文生视频任务状态

POST/kling/v1/videos/image2videoKling 图生视频

GET/kling/v1/videos/image2video/{task_id}获取 Kling 图生视频任务状态

POST/jimeng/即梦视频生成

POST/v1/video/generations创建视频生成任务

GET/v1/video/generations/{task_id}获取视频生成任务状态

Claude format (Messages)

POST/v1/messagesClaude 聊天

Gemini format

POST/v1beta/models/{model}:generateContentGemini 图片(Nano Banana)

POST/v1/engines/{model}/embeddingsGemini 嵌入(Embeddings)

OpenAI format (Embeddings)

POST/v1/embeddings创建文本嵌入

Text completions (Completions)

POST/v1/completions创建文本补全

OpenAI Audio

POST/v1/audio/transcriptions音频转录

POST/v1/audio/translations音频翻译

POST/v1/audio/speech文本转语音

Rerank

POST/v1/rerank文档重排序

Moderations

POST/v1/moderations内容审核

Realtime

GET/v1/realtime实时 WebSocket 连接

Request format

All requests use JSON:

POST /v1/chat/completions HTTP/1.1
Host: hub.lurus.cn
Content-Type: application/json
Authorization: Bearer sk-your-api-key

{
  "model": "deepseek-chat",
  "messages": [
    {"role": "user", "content": "Hello!"}
  ]
}

Response format

For the type of each field and whether it is required, see Data structures / Schema (auto-synced from the OpenAPI spec and consistent with the gateway implementation).

Successful response

{
  "id": "chatcmpl-abc123",
  "object": "chat.completion",
  "created": 1677858242,
  "model": "deepseek-chat",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello! How can I help you today?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 10,
    "completion_tokens": 12,
    "total_tokens": 22
  }
}

Error response

{
  "error": {
    "code": "invalid_api_key",
    "message": "Invalid API key provided",
    "type": "authentication_error"
  }
}

Rate limits

60RPM (requests/minute)

100,000TPM (tokens/minute)

10Max concurrent requests

Limit type	Default	Description
RPM (requests/minute)	60	Maximum requests per minute
TPM (tokens/minute)	100,000	Maximum tokens per minute
Concurrent requests	10	Maximum requests in flight at once

Exceeding a limit returns a 429 Too Many Requests error; for how to handle it, see Error handling.

SDK support

Python (official OpenAI SDK)

pip install openai

from openai import OpenAI

client = OpenAI(
    base_url="https://hub.lurus.cn/v1",
    api_key="sk-your-api-key"
)

Node.js

npm install openai

import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://hub.lurus.cn/v1',
  apiKey: 'sk-your-api-key'
});

Go

import "github.com/sashabaranov/go-openai"

config := openai.DefaultConfig("sk-your-api-key")
config.BaseURL = "https://hub.lurus.cn/v1"
client := openai.NewClientWithConfig(config)

cURL

curl https://hub.lurus.cn/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-your-api-key" \
  -d '{"model": "deepseek-chat", "messages": [{"role": "user", "content": "Hi"}]}'

Special features

Gateway-enhanced capabilities

Streaming responses

Set stream: true to enable Server-Sent Events and return tokens incrementally

Model routing

Automatically routes to the matching upstream channel by model name, with no need to worry about the underlying configuration

Automatic retries

On a failed request, automatically switches to a backup channel (if multiple are configured)

Streaming responses

Set stream: true to enable Server-Sent Events streaming responses:

{
  "model": "deepseek-chat",
  "messages": [...],
  "stream": true
}

For the full streaming data format and token-by-token handling, see Chat Completions — Streaming responses.

Model routing

Automatically routes to the matching upstream channel by model name, with no need to worry about the underlying configuration.

Automatic retries

On a failed request, automatically switches to a backup channel (if multiple are configured).

---

Next steps

Authentication in detailChat CompletionsData structures / SchemaError handlingGet an API Key

Related Products

Commonly paired with

Lurus APIA multi-tenant LLM gateway deployed in the customer's own environment · Based on the open-source New API, AGPLv3

Unified IdentitySign-in and identity shared by Lurus services · Standard OIDC / OAuth2 · Internal foundation