SDK 接入

凡是允许改 base URL 的 OpenAI 兼容客户端都能直接用。下面是常见几种的最小示例,模型名换成模型库里的任意模型即可。models、reasoning 这类扩展字段通过各 SDK 的 extra_body 之类的透传参数传。

curl

POST /v1/chat/completions
curl https://inferstream.cn/v1/chat/completions \
  -H "Authorization: Bearer $INFERSTREAM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-chat",
    "messages": [
      {"role": "system", "content": "你是一个简洁的助手。"},
      {"role": "user", "content": "用一句话介绍杭州。"}
    ]
  }'

流式与其他参数见对话补全。

Python(openai)

pip install openai
import os
from openai import OpenAI

client = OpenAI(
    base_url="https://inferstream.cn/v1",
    api_key=os.environ["INFERSTREAM_API_KEY"],
)

resp = client.chat.completions.create(
    model="deepseek-chat",
    messages=[{"role": "user", "content": "用一句话介绍杭州。"}],
)
print(resp.choices[0].message.content)
print(resp.usage)

base_url / api_key 也可用环境变量 OPENAI_BASE_URL / OPENAI_API_KEY 提供。流式示例见对话补全。

Node.js(openai)

npm install openai
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://inferstream.cn/v1",
  apiKey: process.env.INFERSTREAM_API_KEY,
});

const resp = await client.chat.completions.create({
  model: "deepseek-chat",
  messages: [{ role: "user", content: "用一句话介绍杭州。" }],
});
console.log(resp.choices[0].message.content);

Anthropic SDK

走 Claude 兼容接口:base URL 指到 https://inferstream.cn/claude,模型名不限于 Claude 系列。

pip install anthropic
from anthropic import Anthropic

client = Anthropic(base_url="https://inferstream.cn/claude", api_key="sk-...")

message = client.messages.create(
    model="deepseek-chat",
    max_tokens=1024,
    messages=[{"role": "user", "content": "用一句话介绍杭州。"}],
)
print(message.content[0].text)

Google(google-genai)

走 Gemini 兼容接口:base URL 指到 https://inferstream.cn/gemini。

pip install google-genai(需自行验证)
from google import genai
from google.genai import types

client = genai.Client(
    api_key="sk-...",
    http_options=types.HttpOptions(base_url="https://inferstream.cn/gemini"),
)
response = client.models.generate_content(
    model="gemini-2.5-flash",
    contents="用一句话介绍杭州。",
)
print(response.text)
注意

未逐版本验证:google-genai 的 http_options.base_url 写法未在本站逐版本验证,若 SDK 版本不同请对照其文档调整。

LangChain

pip install langchain-openai
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    model="deepseek-chat",
    base_url="https://inferstream.cn/v1",
    api_key="sk-...",
)
print(llm.invoke("用一句话介绍杭州。").content)

ChatOpenAI 的 base_url / api_key 也可用环境变量 OPENAI_BASE_URL / OPENAI_API_KEY 提供。models、reasoning 这类扩展字段通过 extra_body 传。

Vercel AI SDK

npm install ai @ai-sdk/openai
import { createOpenAI } from "@ai-sdk/openai";
import { generateText } from "ai";

const inferstream = createOpenAI({
  baseURL: "https://inferstream.cn/v1",
  apiKey: process.env.INFERSTREAM_API_KEY,
});

// 用 .chat() 走 /v1/chat/completions;不带 .chat() 时 AI SDK 5 默认走 /v1/responses
const { text } = await generateText({
  model: inferstream.chat("deepseek-chat"),
  prompt: "用一句话介绍杭州。",
});
console.log(text);

用 .chat() 走 /v1/chat/completions;不带 .chat() 时 AI SDK 5 默认走 /v1/responses,两条接口都开放,区别见 Responses。