OpenAI SDK 调用

CloudBase AI 兼容 OpenAI API 协议，可直接使用 OpenAI SDK 调用，方便项目迁移和多模型切换。

前置条件

1. 已开通云开发环境
2. 已配置大模型（见 大模型配置指南）
3. 已创建 API Key（获取地址）

安装

# Node.js
npm install openai

# Python
pip install openai

配置

将 OpenAI SDK 的 baseURL 和 apiKey 替换为 CloudBase 的配置：

配置项	值
baseURL	https://<ENV_ID>.api.tcloudbasegateway.com/v1/ai/<PROVIDER>/v1
apiKey	你的 CloudBase API Key

支持的 Provider

Provider	说明
hunyuan	腾讯混元
deepseek	DeepSeek

Node.js 示例

非流式调用

const OpenAI = require("openai");

const client = new OpenAI({
  apiKey: "<YOUR_API_KEY>",
  baseURL: "https://<ENV_ID>.api.tcloudbasegateway.com/v1/ai/deepseek/v1"
});

async function main() {
  const completion = await client.chat.completions.create({
    model: "deepseek-r1-0528",
    messages: [{ role: "user", content: "你好" }],
    temperature: 0.7
  });

  console.log(completion.choices[0].message.content);
}

main();

流式调用

const OpenAI = require("openai");

const client = new OpenAI({
  apiKey: "<YOUR_API_KEY>",
  baseURL: "https://<ENV_ID>.api.tcloudbasegateway.com/v1/ai/deepseek/v1"
});

async function main() {
  const stream = await client.chat.completions.create({
    model: "deepseek-r1-0528",
    messages: [{ role: "user", content: "介绍一下你自己" }],
    stream: true
  });

  for await (const chunk of stream) {
    const content = chunk.choices[0]?.delta?.content || "";
    process.stdout.write(content);
  }
}

main();

Python 示例

非流式调用

from openai import OpenAI

client = OpenAI(
    api_key="<YOUR_API_KEY>",
    base_url="https://<ENV_ID>.api.tcloudbasegateway.com/v1/ai/deepseek/v1"
)

completion = client.chat.completions.create(
    model="deepseek-r1-0528",
    messages=[{"role": "user", "content": "你好"}],
    temperature=0.7
)

print(completion.choices[0].message.content)

流式调用

from openai import OpenAI

client = OpenAI(
    api_key="<YOUR_API_KEY>",
    base_url="https://<ENV_ID>.api.tcloudbasegateway.com/v1/ai/deepseek/v1"
)

stream = client.chat.completions.create(
    model="deepseek-r1-0528",
    messages=[{"role": "user", "content": "介绍一下你自己"}],
    stream=True
)

for chunk in stream:
    content = chunk.choices[0].delta.content or ""
    print(content, end="", flush=True)

多轮对话

const OpenAI = require("openai");

const client = new OpenAI({
  apiKey: "<YOUR_API_KEY>",
  baseURL: "https://<ENV_ID>.api.tcloudbasegateway.com/v1/ai/hunyuan/v1"
});

const messages = [
  { role: "system", content: "你是一个诗词专家" }
];

async function chat(userMessage) {
  messages.push({ role: "user", content: userMessage });

  const completion = await client.chat.completions.create({
    model: "hunyuan-turbos-latest",
    messages
  });

  const assistantMessage = completion.choices[0].message;
  messages.push(assistantMessage);

  return assistantMessage.content;
}

// 使用示例
await chat("李白最著名的诗是什么？");
await chat("这首诗的创作背景是什么？");

切换模型

只需修改 baseURL 中的 provider 和 model 参数即可切换不同模型：

// 使用混元
const hunyuanClient = new OpenAI({
  apiKey: "<YOUR_API_KEY>",
  baseURL: "https://<ENV_ID>.api.tcloudbasegateway.com/v1/ai/hunyuan/v1"
});

await hunyuanClient.chat.completions.create({
  model: "hunyuan-turbos-latest",
  messages: [{ role: "user", content: "你好" }]
});

// 使用 DeepSeek
const deepseekClient = new OpenAI({
  apiKey: "<YOUR_API_KEY>",
  baseURL: "https://<ENV_ID>.api.tcloudbasegateway.com/v1/ai/deepseek/v1"
});

await deepseekClient.chat.completions.create({
  model: "deepseek-r1-0528",
  messages: [{ role: "user", content: "你好" }]
});

支持的参数

参数	类型	说明
model	string	模型名称
messages	array	消息列表
stream	boolean	是否流式返回
temperature	number	采样温度 (0-2)
top_p	number	核采样 (0-1)
max_tokens	number	最大生成 token 数
presence_penalty	number	存在惩罚
frequency_penalty	number	频率惩罚

从 OpenAI 迁移

如果你的项目原本使用 OpenAI，迁移到 CloudBase 只需两步：

1. 修改配置

// 原 OpenAI 配置
const client = new OpenAI({
  apiKey: "sk-xxx"  // OpenAI API Key
});

// 改为 CloudBase 配置
const client = new OpenAI({
  apiKey: "<YOUR_CLOUDBASE_API_KEY>",
  baseURL: "https://<ENV_ID>.api.tcloudbasegateway.com/v1/ai/deepseek/v1"
});

2. 修改模型名称

// 原 OpenAI 模型
model: "gpt-4"

// 改为 CloudBase 支持的模型
model: "deepseek-r1-0528"  // 或 hunyuan-turbos-latest 等

错误处理

const OpenAI = require("openai");

const client = new OpenAI({
  apiKey: "<YOUR_API_KEY>",
  baseURL: "https://<ENV_ID>.api.tcloudbasegateway.com/v1/ai/deepseek/v1"
});

try {
  const completion = await client.chat.completions.create({
    model: "deepseek-r1-0528",
    messages: [{ role: "user", content: "你好" }]
  });
  console.log(completion.choices[0].message.content);
} catch (error) {
  if (error instanceof OpenAI.APIError) {
    console.error("API 错误:", error.status, error.message);
  } else {
    throw error;
  }
}

相关文档

• 概述
• cURL 调用
• 大模型配置指南