AI 调度台

统一分发 API Key，追踪每一次调用，
清楚看见团队的 AI 成本与流向。

✣Claude

⬡Codex

◎OpenAI

✦Gemini

◇国产模型

request.http

1curl -X POST https://api.yunheliai.com/v1/chat/completions \2-H "Content-Type: application/json" \3-H "Authorization: Bearer sk-••••••••" \4-d '{5  "base_url": "https://api.yunheliai.com/v1",6  "model": "gpt-5.5",7  "messages": [8    {"role": "system", "content": "你是运河里 AI 调度台助手"},9    {"role": "user", "content": "请帮我生成接口调用示例"}10  ],11  "stream": true,12  "temperature": 0.7,13  "max_tokens": 204814}'

response.json

200 OK示例响应1.42s

{
  "id": "chatcmpl-demo",
  "object": "chat.completion",
  "usage": {
    "prompt_tokens": 129,
    "completion_tokens": 256,
    "total_tokens": 385
  }
}

⬡Codex API route

200 OK

POST /v1/chat/completions
Authorization: Bearer sk-••••••••
model: gpt-5.5
id: chatcmpl-demo

prompt 129output 2561.42s

统一分发

为每位成员分配独立 API Key

实时追踪

记录模型、Token、时间与调用来源

成本流向

看清每个团队与模型的实际花费