Skip to main content
TorchV文档中心

开放平台应用

开放平台模块提供标准化的开放API接口，包括Chat Completion对话接口和开放平台应用管理接口。

Chat Completion V1​

接口地址​

POST /openapi/api/v1/chat/completions

请求方法​

POST

权限要求​

chat_v1 READ

接口描述​

标准的Chat Completion接口，支持SSE流式输出和JSON响应。

请求Header​

Header名类型必填说明
AuthorizationString是Bearer Token认证，格式：Bearer sk_xxxxxxxxxxxxx

请求参数​

参数名类型必填默认值说明
userIdString否-用户ID
conversationIdString否-会话ID
promptString否-提示词文本
promptTemplateString否-提示词模板
modelString否-模型名称
ragSearchboolean否true是否启用RAG检索
reasonStatusboolean否false是否返回推理状态
messagesLinkedList<BotMessage>否-消息列表
containerIdList<String>否-知识库容器ID列表
multipleDialogboolean否true是否多轮对话
temperaturefloat否0.95温度参数，0-1
top_pfloat否0.7Top P采样
fileIdsList<String>否-文件ID列表
apiKeyString否-API Key（隐藏字段）

BotMessage结构：

参数名类型必填说明
roleString是角色类型
contentObject否消息内容
reasoningContentString否推理内容

RequestBody示例​

{
  "userId": "user_001",
  "conversationId": "conv_001",
  "prompt": "你好",
  "model": "glm-4",
  "ragSearch": true,
  "reasonStatus": false,
  "messages": [
    {
      "role": "user",
      "content": "你好，请介绍一下自己"
    }
  ],
  "containerId": ["container_001"],
  "multipleDialog": true,
  "temperature": 0.95,
  "top_p": 0.7,
  "fileIds": []
}

响应示例​

SSE流式或JSON响应：

{
  "id": "chatcmpl-123",
  "object": "chat.completion",
  "created": 1677652288,
  "model": "glm-4",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "你好！我是智能助手，很高兴为您服务。"
      },
      "finish_reason": "stop"
    }
  ]
}

请求示例​

curl -X POST "${domain}/openapi/api/v1/chat/completions" \
  -H "Authorization: Bearer sk_xxxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-4",
    "prompt": "你好",
    "messages": [
      {"role": "user", "content": "你好，请介绍一下自己"}
    ],
    "containerId": ["container_001"],
    "temperature": 0.95,
    "top_p": 0.7
  }'

查询可用模型列表​

接口地址​

GET /openapi/community/v1/models

请求方法​

GET

权限要求​

chat_v2 READ

接口描述​

查询当前可用的模型列表，返回OpenAI兼容格式。

请求参数​

无

响应示例​

{
  "object": "list",
  "data": [
    {
      "id": "glm-4",
      "object": "model",
      "created": 1677652288,
      "owned_by": "zhipu"
    },
    {
      "id": "glm-3-turbo",
      "object": "model",
      "created": 1677652288,
      "owned_by": "zhipu"
    }
  ]
}

请求示例​

curl -X GET "${domain}/openapi/community/v1/models" \
  -H "Authorization: Bearer sk_xxxxxxxxxxxxx" \
  -H "Content-Type: application/json"

Chat OpenAI兼容接口​

接口地址​

POST /openapi/community/v1/chat/completions

请求方法​

POST

权限要求​

chat_v2 READ

接口描述​

OpenAI兼容的Chat Completion接口V2，支持SSE流式输出和JSON响应。

请求Header​

Header名类型必填说明
AuthorizationString是Bearer Token认证，格式：Bearer sk_xxxxxxxxxxxxx

请求参数​

参数名类型必填默认值说明
messagesLinkedList<BotMessage>是-消息列表
modelString否-模型名称
temperaturefloat否0.95温度参数，0-1
top_pfloat否0.7Top P采样
conversationIdString否-会话ID
containerIdList<String>否-知识库容器ID列表
userString否-用户标识

BotMessage结构：

参数名类型必填说明
roleString是角色类型
contentObject否消息内容
reasoningContentString否推理内容

RequestBody示例​

{
  "messages": [
    {
      "role": "user",
      "content": "你好，请介绍一下自己"
    }
  ],
  "model": "glm-4",
  "temperature": 0.95,
  "top_p": 0.7,
  "conversationId": "conv_001",
  "containerId": ["container_001"],
  "user": "user_001"
}

响应示例​

SSE流式或JSON响应：

{
  "id": "chatcmpl-123",
  "object": "chat.completion",
  "created": 1677652288,
  "model": "glm-4",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "你好！我是智能助手，很高兴为您服务。"
      },
      "finish_reason": "stop"
    }
  ]
}

请求示例​

curl -X POST "${domain}/openapi/community/v1/chat/completions" \
  -H "Authorization: Bearer sk_xxxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-4",
    "messages": [
      {"role": "user", "content": "你好，请介绍一下自己"}
    ],
    "containerId": ["container_001"],
    "temperature": 0.95,
    "top_p": 0.7
  }'

应用面板信息​

接口地址​

GET /openapi/store/api/v1/dashboard

请求方法​

GET

权限要求​

app_dashboard READ

接口描述​

获取应用面板配置信息。

请求Header​

Header名类型必填说明
AuthorizationString是认证Token

请求参数​

无

响应示例​

{
  "code": 8200,
  "message": "操作成功",
  "data": {
    "id": 1,
    "sort": 1,
    "tenantId": "0",
    "code": "app_001",
    "title": "智能客服",
    "description": "智能客服应用",
    "logoImage": "https://cdn.example.com/logo.png",
    "appUrl": "https://app.example.com/app_001"
  }
}

请求示例​

curl -X GET "${domain}/openapi/store/api/v1/dashboard" \
  -H "Authorization: Bearer sk_xxxxxxxxxxxxx" \
  -H "Content-Type: application/json"

按AppID查询应用​

接口地址​

GET /openapi/store/api/v1/queryByAppId

请求方法​

GET

权限要求​

app_info READ

接口描述​

根据AppID查询应用信息。

请求参数​

参数名类型必填说明
appidString是应用ID（注意：参数名为小写appid）

响应示例​

{
  "code": 8200,
  "message": "操作成功",
  "data": {
    "appId": "app_001",
    "name": "智能客服",
    "description": "智能客服应用",
    "icon": "https://cdn.example.com/icon.png",
    "status": "published",
    "config": {
      "llmModel": "glm-4",
      "llmTemperature": 0.7
    }
  }
}

请求示例​

curl -X GET "${domain}/openapi/store/api/v1/queryByAppId?appid=app_001" \
  -H "Authorization: Bearer sk_xxxxxxxxxxxxx" \
  -H "Content-Type: application/json"

应用商店列表​

接口地址​

GET /openapi/store/api/v1/store/list

请求方法​

GET

权限要求​

app_dashboard READ

接口描述​

分页查询应用市场列表。

请求参数​

参数名类型必填说明
pageNoInteger是当前页码，默认1
pageSizeInteger是页码大小，默认10
nameString否应用名称（模糊查询）

响应示例​

{
  "code": 8200,
  "message": "操作成功",
  "data": [
    {
      "id": 1,
      "appid": "app_001",
      "role": "assistant",
      "category": "客服",
      "icon": "https://cdn.example.com/icon.png",
      "name": "智能客服",
      "description": "智能客服应用",
      "accessToken": "token_xxx",
      "appUrl": "https://app.example.com/app_001",
      "tag": "AI,客服",
      "creator": "admin",
      "createTime": "2024-01-01 10:00:00",
      "modifier": "admin",
      "modifierTime": "2024-01-01 10:00:00",
      "sort": 1
    }
  ],
  "total": 1,
  "totalPage": 1,
  "pageNo": 1,
  "pageSize": 10
}

请求示例​

curl -X GET "${domain}/openapi/store/api/v1/store/list?pageNo=1&pageSize=10&name=客服" \
  -H "Authorization: Bearer sk_xxxxxxxxxxxxx" \
  -H "Content-Type: application/json"

提交面板反馈​

接口地址​

POST /openapi/store/feedback/api/v1/add

请求方法​

POST

权限要求​

app_chat_feedback WRITE

接口描述​

提交应用面板反馈。

请求参数​

参数名类型必填说明
conversationIdString是会话ID
modelString否模型名称，默认glm-3-turbo
questionString是问题内容
answerString是回答内容
remarkString否备注信息
typeString是反馈类型（like/tap/other）
tagString否标签
messageIdString否消息ID
appidString否应用ID

RequestBody示例​

{
  "conversationId": "conv_001",
  "model": "glm-3-turbo",
  "question": "什么是知识库？",
  "answer": "知识库是用于存储和管理知识的系统...",
  "remark": "",
  "type": "like",
  "tag": "",
  "messageId": "msg_001",
  "appid": "app_001"
}

响应示例​

{
  "code": 8200,
  "message": "提交成功",
  "data": "feedback_001"
}

请求示例​

curl -X POST "${domain}/openapi/store/feedback/api/v1/add" \
  -H "Authorization: Bearer sk_xxxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "conversationId": "conv_001",
    "question": "什么是知识库？",
    "answer": "知识库是用于存储和管理知识的系统...",
    "type": "like",
    "appid": "app_001"
  }'
Previous
应用管理
Next
知识库