创建聊天补全。与 OpenAI Chat Completions API 兼容。

Create a chat completion. Compatible with the OpenAI Chat Completions API.

请求参数

Request Parameters

示例（非流式）

Example (non-streaming)

curl http://<server>:9090/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $API_KEY" \
  -d '{
    "model": "Qwen3.6",
    "messages": [{"role": "user", "content": "你好"}],
    "max_tokens": 100
  }'

from openai import OpenAI

client = OpenAI(
    base_url="http://<server>:9090/v1",
    api_key="llm-xxx"
)
response = client.chat.completions.create(
    model="Qwen3.6",
    messages=[{"role": "user", "content": "你好"}],
    max_tokens=100
)
print(response.choices[0].message.content)

const resp = await fetch("http://<server>:9090/v1/chat/completions", {
    method: "POST",
    headers: {
        "Content-Type": "application/json",
        "Authorization": "Bearer llm-xxx"
    },
    body: JSON.stringify({
        model: "Qwen3.6",
        messages: [{role: "user", content: "你好"}],
        max_tokens: 100
    })
});
const data = await resp.json();
console.log(data.choices[0].message.content);

流式响应

Streaming Response

设置 stream: true 后，响应以 SSE 格式返回：

When stream: true, the response is returned as SSE:

data: {"choices":[{"delta":{"role":"assistant","content":null}}],...}
data: {"choices":[{"delta":{"content":"你好"}}],...}
data: {"choices":[{"delta":{"content":"！"}}],...}
...
data: [DONE]

每条 data: 行包含一个 JSON 对象，delta.content 为增量文本。流以 data: [DONE] 结束。

Each data: line contains a JSON object with delta.content as incremental text. The stream ends with data: [DONE].

响应格式

Response Format

{
  "id": "chatcmpl-xxx",
  "object": "chat.completion",
  "created": 1784790525,
  "model": "Qwen3.6",
  "choices": [{
    "index": 0,
    "message": { "role": "assistant", "content": "你好！有什么可以帮助你的？" },
    "finish_reason": "stop"
  }],
  "usage": { "prompt_tokens": 13, "completion_tokens": 10, "total_tokens": 23 }
}

Anthropic Messages API 兼容接口。与 Claude API 格式一致。

Anthropic-compatible Messages API. Same format as the Claude API.

认证

Authentication

使用 x-api-key 请求头（也支持 Authorization: Bearer）：

Use the x-api-key header (also supports Authorization: Bearer):

curl -X POST http://<server>:9090/v1/messages \
  -H "x-api-key: llm-xxx" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "Qwen3.6",
    "messages": [{"role": "user", "content": "你好"}],
    "max_tokens": 100
  }'

请求参数

Request Parameters

响应格式

Response Format

{
  "id": "msg_...",
  "type": "message",
  "role": "assistant",
  "content": [{"type": "text", "text": "你好！有什么可以帮助你的？"}],
  "model": "Qwen3.6",
  "stop_reason": "end_turn",
  "usage": {"input_tokens": 13, "output_tokens": 14}
}

Python SDK 示例

Python SDK Example

import anthropic

client = anthropic.Anthropic(
    base_url="http://<server>:9090/v1",
    api_key="llm-xxx",
)
response = client.messages.create(
    model="Qwen3.6",
    messages=[{"role": "user", "content": "你好"}],
    max_tokens=100,
)
print(response.content[0].text)

列出所有可用模型。

List all available models.

示例

Example

curl http://<server>:9090/v1/models -H "Authorization: Bearer $API_KEY"

响应

Response

{
  "object": "list",
  "data": [{
    "id": "Qwen3.6",
    "object": "model",
    "owned_by": "llamacpp-qwen",
    "description": "Qwen3.6-27B (llama.cpp)"
  }]
}

创建 API Key。需要 JWT Cookie 认证。Key 仅在创建时返回一次，请妥善保存。

Create an API Key. Requires JWT Cookie auth. The key is shown only once — save it securely.

请求

Request

curl http://<server>:9090/api/keys/create \
  -H "Content-Type: application/json" \
  -b "llm_gateway_token=$TOKEN" \
  -d '{"name": "my-app-key"}'

响应

Response

{
  "name": "my-app-key",
  "key": "llm-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  "prefix": "llm-xxxxxxxx"
}

列出当前用户的所有 API Key（不包含完整 Key 值）。

List all API Keys for the current user (full key value not included).

响应

Response

[{
  "id": 1, "user_id": 1, "name": "my-app-key",
  "key_prefix": "llm-xxxxxxxx", "usage_count": 42,
  "last_used_at": "2026-07-23 08:30:00",
  "created_at": "2026-07-22 12:00:00"
}]

删除指定的 API Key。删除后使用该 Key 的客户端将无法调用 API。

Delete the specified API Key. Clients using this key will lose API access.

curl -X DELETE http://<server>:9090/api/keys/1 -b "llm_gateway_token=$TOKEN"