快速开始 | JoyToken | Documentation

前提条件

项目	值
API Key	`JOY_TOKEN_API_KEY`
API Base URL	从环境选择 `JOY_TOKEN_API_BASE_URL`
OpenAI Base URL	从环境选择 `JOY_TOKEN_OPENAI_BASE_URL`
模型	`auto`
TypeScript	Node.js `18+`
Go	Go `1.22+`

第 1 步：设置环境

$ export JOY_TOKEN_API_KEY="jt_xxx"
$ export JOY_TOKEN_API_BASE_URL="https://api-dev.joytokens.ai"
$ export JOY_TOKEN_OPENAI_BASE_URL="https://api-dev.joytokens.ai/openai/v1"

第 2 步：安装 SDK

TypeScript:

$ npm install @joytoken/client-sdk-ts

Go:

$ go get github.com/joytoken/client-sdk-golang

第 3 步：发起调用

TypeScript:

1 import { JoyTokenClient } from "@joytoken/client-sdk-ts";
2 
3 const joytoken = new JoyTokenClient({
4   apiKey: process.env.JOY_TOKEN_API_KEY,
5   apiBaseUrl: process.env.JOY_TOKEN_API_BASE_URL,
6   openAIBaseUrl: process.env.JOY_TOKEN_OPENAI_BASE_URL,
7 });
8 
9 const completion = await joytoken.chat.completions.create({
10   model: "auto",
11   messages: [{ role: "user", content: "Reply with exactly: pong" }],
12 });
13 
14 console.log(completion.choices[0]?.message?.content);

Go:

1 client := joytoken.NewClient(
2 	joytoken.WithAPIKey(os.Getenv("JOY_TOKEN_API_KEY")),
3 	joytoken.WithAPIBaseURL(os.Getenv("JOY_TOKEN_API_BASE_URL")),
4 	joytoken.WithOpenAIBaseURL(os.Getenv("JOY_TOKEN_OPENAI_BASE_URL")),
5 )
6 
7 completion, err := client.CreateChatCompletion(ctx, joytoken.ChatCompletionRequest{
8 	Model: "auto",
9 	Messages: []joytoken.ChatMessage{
10 		{Role: "user", Content: "Reply with exactly: pong"},
11 	},
12 })
13 if err != nil {
14 	return err
15 }
16 fmt.Println(completion.Choices[0].Message.Content)

第 4 步：检查结果

位置	应看到
程序输出	`pong`
JoyToken Logs	请求路径是 `/openai/v1/chat/completions`
JoyToken Usage	本次调用产生 tokens 和 credits

常见错误

错误	处理方式
`401 Unauthorized`	检查 `JOY_TOKEN_API_KEY` 是否存在且完整
`403 Forbidden`	检查 API Key 策略是否允许 `auto` 或目标模型
`402 insufficient_quota`	检查钱包余额和 Key 预算
请求超时	设置更大的 SDK timeout，或检查到所选环境的网络