SDK 与 Fetch

把 ttttt.ai 接进你自己的应用 / 后端服务。绝大多数 SDK 只需要改两项：base_url 和 api_key。

OpenAI SDK

Python


from openai import OpenAI
 
client = OpenAI(
    api_key="owo-...",
    base_url="https://api.ttttt.ai/v1",
)
 
# Chat Completions
resp = client.chat.completions.create(
    model="gpt-5.5",
    messages=[{"role": "user", "content": "你好"}],
)
print(resp.choices[0].message.content)
 
# Streaming
stream = client.chat.completions.create(
    model="gpt-5.5",
    messages=[{"role": "user", "content": "讲个冷笑话"}],
    stream=True,
)
for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="", flush=True)
 
# Embeddings
emb = client.embeddings.create(
    model="text-embedding-3-large",
    input="ttttt.ai 是统一的 AI API 网关。",
)
vec = emb.data[0].embedding

Node.js / TypeScript


import OpenAI from "openai";
 
const client = new OpenAI({
  apiKey: process.env.TTTTT_API_KEY,           // owo-...
  baseURL: "https://api.ttttt.ai/v1",
});
 
const resp = await client.chat.completions.create({
  model: "gpt-5.5",
  messages: [{ role: "user", content: "你好" }],
});
console.log(resp.choices[0].message.content);

通过 Responses API

新接口推荐：


resp = client.responses.create(
    model="gpt-5.5",
    input="用一句话解释 transformer。",
)
print(resp.output_text)
 
# 多轮
follow = client.responses.create(
    model="gpt-5.5",
    previous_response_id=resp.id,
    input="再展开一点。",
)

Anthropic SDK

Python


import anthropic
 
client = anthropic.Anthropic(
    api_key="owo-...",
    base_url="https://api.ttttt.ai",
)
 
msg = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "你好"}],
)
print(msg.content[0].text)

Node.js / TypeScript


import Anthropic from "@anthropic-ai/sdk";
 
const client = new Anthropic({
  apiKey: process.env.TTTTT_API_KEY,
  baseURL: "https://api.ttttt.ai",
});
 
const msg = await client.messages.create({
  model: "claude-sonnet-4-6",
  max_tokens: 1024,
  messages: [{ role: "user", content: "你好" }],
});
console.log(msg.content[0].type === "text" ? msg.content[0].text : "");

原生 fetch / curl

不依赖 SDK 也完全可以——HTTP + JSON：


const r = await fetch("https://api.ttttt.ai/v1/chat/completions", {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.TTTTT_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "gpt-5.5",
    messages: [{ role: "user", content: "ping" }],
  }),
});
const data = await r.json();
console.log(data.choices[0].message.content);


import os, requests
 
r = requests.post(
    "https://api.ttttt.ai/v1/chat/completions",
    headers={
        "Authorization": f"Bearer {os.environ['TTTTT_API_KEY']}",
        "Content-Type": "application/json",
    },
    json={
        "model": "gpt-5.5",
        "messages": [{"role": "user", "content": "ping"}],
    },
)
print(r.json()["choices"][0]["message"]["content"])

LangChain


from langchain_openai import ChatOpenAI
 
llm = ChatOpenAI(
    model="gpt-5.5",
    api_key="owo-...",
    base_url="https://api.ttttt.ai/v1",
    temperature=0.2,
)
print(llm.invoke("用一句话解释 RAG").content)


import { ChatOpenAI } from "@langchain/openai";
 
const llm = new ChatOpenAI({
  model: "gpt-5.5",
  apiKey: process.env.TTTTT_API_KEY,
  configuration: { baseURL: "https://api.ttttt.ai/v1" },
});
const out = await llm.invoke("你好");
console.log(out.content);

LlamaIndex


from llama_index.llms.openai_like import OpenAILike
 
llm = OpenAILike(
    model="gpt-5.5",
    api_base="https://api.ttttt.ai/v1",
    api_key="owo-...",
    is_chat_model=True,
)
print(llm.complete("ping").text)

Vercel AI SDK


import { createOpenAI } from "@ai-sdk/openai";
import { generateText } from "ai";
 
const ttttt = createOpenAI({
  apiKey: process.env.TTTTT_API_KEY,
  baseURL: "https://api.ttttt.ai/v1",
});
 
const { text } = await generateText({
  model: ttttt("gpt-5.5"),
  prompt: "你好",
});
console.log(text);

Go


package main
 
import (
    "context"
    "fmt"
 
    "github.com/sashabaranov/go-openai"
)
 
func main() {
    cfg := openai.DefaultConfig("owo-...")
    cfg.BaseURL = "https://api.ttttt.ai/v1"
    client := openai.NewClientWithConfig(cfg)
 
    resp, _ := client.CreateChatCompletion(
        context.Background(),
        openai.ChatCompletionRequest{
            Model: "gpt-5.5",
            Messages: []openai.ChatCompletionMessage{
                {Role: "user", Content: "你好"},
            },
        },
    )
    fmt.Println(resp.Choices[0].Message.Content)
}

Rust


use async_openai::{config::OpenAIConfig, Client, types::CreateChatCompletionRequestArgs};
 
#[tokio::main]
async fn main() -> anyhow::Result<()> {
    let config = OpenAIConfig::new()
        .with_api_key("owo-...")
        .with_api_base("https://api.ttttt.ai/v1");
    let client = Client::with_config(config);
 
    let req = CreateChatCompletionRequestArgs::default()
        .model("gpt-5.5")
        .messages([("user", "你好").into()])
        .build()?;
    let resp = client.chat().create(req).await?;
    println!("{}", resp.choices[0].message.content.clone().unwrap_or_default());
    Ok(())
}

服务端最佳实践

不要把密钥下发到浏览器 / APP；前端通过你的服务端代理一层。
设超时——SDK 默认超时通常是 60-120s，根据业务调整；流式可以更长。
开启重试——max_retries=2 起步；详见错误码与重试。
记 X-Request-ID——日志里把它和你的业务 trace 关联，工单沟通效率会高很多。
预算护栏——后台任务用专门的低额度密钥 + 低 RPM 上限，避免循环 bug 失控。
fallback 策略——对延迟敏感的接口可以考虑两个 model 串联：先快后慢，先 gpt-5.4 不行再 gpt-5.5。

排错路径

调不通时按下面顺序检查：

curl 一下鉴权页里的最小请求是否 200 —— 如果连 curl 都不通，问题在密钥 / 网络，不是 SDK。
SDK 是不是 base URL 拼成了 https://api.ttttt.ai（没带 /v1）—— OpenAI SDK 必须是 /v1 结尾。
模型 ID 是否正确 —— GET /v1/models 列出当前密钥可用的所有 ID。
看 X-Request-ID 在控制台用量页是否能查到 —— 能查到 = 请求到了 ttttt.ai；查不到 = 客户端根本没发出。