AI 模型接口

• 基础信息
• 二级页面总览
• 通用错误
• 音频 Audio
  • 原生 Gemini 格式
  • 原生 OpenAI 格式
• 聊天 Chat
  • 原生 Claude 格式
  • 原生 Gemini 格式
  • 原生 OpenAI 格式
• 补全 Completions
• 嵌入 Embeddings
  • 原生 OpenAI 格式
  • 原生 Gemini 格式
• 图像 Images
  • 原生 Gemini 格式
  • OpenAI 聊天格式
  • 原生 OpenAI 格式
  • 通义千问 OpenAI 格式
• 模型 Models
• 审查 Moderations
• 实时语音 Realtime
• 重排序 Rerank

• Base URL: https://api.infillm.com
• 鉴权方式: Authorization: Bearer <API_KEY>
• 兼容鉴权: Gemini 原生格式也可使用 ?key=<API_KEY>；Claude 原生格式也可使用 x-api-key: <API_KEY>
• 请求格式: JSON；音频转录、音频翻译等文件上传接口使用 multipart/form-data
• 流式输出: 支持流式的接口可传入 stream: true，Gemini 流式接口可使用 :streamGenerateContent?alt=sse

{
  "error": {
    "message": "错误信息",
    "type": "invalid_request_error",
    "code": "invalid_api_key"
  }
}

curl "https://api.infillm.com/v1beta/models/gemini-2.5-flash-preview-tts:generateContent" \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "role": "user",
        "parts": [
          {"text": "请用自然的语气朗读：你好，欢迎使用 InfillM API。"}
        ]
      }
    ],
    "generationConfig": {
      "responseModalities": ["AUDIO"],
      "speechConfig": {
        "voiceConfig": {
          "prebuiltVoiceConfig": {
            "voiceName": "Kore"
          }
        }
      }
    }
  }'

curl https://api.infillm.com/v1/audio/speech \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -o speech.mp3 \
  -d '{
    "model": "tts-1",
    "input": "你好，欢迎使用 InfillM API。",
    "voice": "alloy",
    "response_format": "mp3"
  }'

curl https://api.infillm.com/v1/audio/transcriptions \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -F file="@speech.mp3" \
  -F model="whisper-1" \
  -F language="zh" \
  -F response_format="json"

curl https://api.infillm.com/v1/audio/translations \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -F file="@speech.mp3" \
  -F model="whisper-1"

curl https://api.infillm.com/v1/messages \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-3-5-sonnet-20241022",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "用一句话解释 API 网关。"}
    ]
  }'

curl "https://api.infillm.com/v1beta/models/gemini-2.5-pro:generateContent" \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "role": "user",
        "parts": [
          {"text": "用一句话介绍向量数据库。"}
        ]
      }
    ],
    "generationConfig": {
      "temperature": 0.7,
      "maxOutputTokens": 512
    }
  }'

{
  "contents": [
    {
      "role": "user",
      "parts": [
        {"text": "请描述这张图片。"},
        {
          "inlineData": {
            "mimeType": "image/png",
            "data": "<base64-image>"
          }
        }
      ]
    }
  ]
}

curl https://api.infillm.com/v1/chat/completions \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [
      {"role": "system", "content": "你是一个简洁的中文助手。"},
      {"role": "user", "content": "用一句话介绍向量数据库。"}
    ],
    "temperature": 0.7
  }'

curl https://api.infillm.com/v1/responses \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "input": "总结一下 API 网关的作用。"
  }'

curl https://api.infillm.com/v1/responses/compact \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "input": [
      {"role": "user", "content": "这里是一段很长的历史对话..."}
    ]
  }'

curl https://api.infillm.com/v1/completions \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-3.5-turbo-instruct",
    "prompt": "把下面这句话改写得更专业：我们会很快处理你的问题。",
    "max_tokens": 128,
    "temperature": 0.3
  }'

curl https://api.infillm.com/v1/embeddings \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-embedding-3-small",
    "input": [
      "向量数据库用于语义检索。",
      "API 网关可以统一鉴权、计费和限流。"
    ]
  }'

curl https://api.infillm.com/v1/engines/text-embedding-3-small/embeddings \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-embedding-3-small",
    "input": "API 网关统一处理鉴权、限流和计费。"
  }'

curl "https://api.infillm.com/v1beta/models/gemini-2.5-flash-image:generateContent" \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      {
        "role": "user",
        "parts": [
          {"text": "draw a cat"}
        ]
      }
    ],
    "generationConfig": {
      "responseModalities": ["TEXT", "IMAGE"],
      "imageConfig": {
        "aspectRatio": "16:9",
        "imageSize": "4K"
      }
    }
  }'

curl https://api.infillm.com/v1/chat/completions \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-2.5-flash-image",
    "stream": true,
    "messages": [
      {"role": "user", "content": "draw a cat"}
    ],
    "contents": [
      {
        "role": "user",
        "parts": [
          {"text": "draw a cat"}
        ]
      }
    ]
  }'

curl https://api.infillm.com/v1/images/generations \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-1",
    "prompt": "一张极简产品海报，白色背景，玻璃质感的智能音箱，柔和自然光。",
    "size": "1024x1024"
  }'

curl https://api.infillm.com/v1/images/edits \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -F image="@image.png" \
  -F model="gpt-image-1" \
  -F prompt="A cute baby sea otter wearing a beret." \
  -F size="1024x1024"

curl https://api.infillm.com/v1/images/generations \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen-image-plus",
    "input": {
      "messages": [
        {
          "role": "user",
          "content": [
            {"text": "一张极简产品海报，白色背景，玻璃质感的智能音箱，柔和自然光。"}
          ]
        }
      ]
    },
    "parameters": {
      "size": "1328*1328",
      "prompt_extend": true,
      "watermark": false
    }
  }'

curl https://api.infillm.com/v1/images/edits \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen-image-edit-plus",
    "input": {
      "messages": [
        {
          "role": "user",
          "content": [
            {"image": "https://example.com/input.png"},
            {"text": "将背景替换为清晨的城市街道，保留主体。"}
          ]
        }
      ]
    },
    "parameters": {
      "n": 1,
      "prompt_extend": true,
      "watermark": false
    }
  }'

curl https://api.infillm.com/v1/models \
  -H "Authorization: Bearer $INFILLM_API_KEY"

{
  "object": "list",
  "data": [
    {
      "id": "gpt-4o-mini",
      "object": "model",
      "created": 1710000000,
      "owned_by": "openai"
    }
  ]
}

curl https://api.infillm.com/v1beta/models \
  -H "Authorization: Bearer $INFILLM_API_KEY"

{
  "models": [
    {
      "name": "models/gemini-pro",
      "version": "string",
      "displayName": "string",
      "description": "string",
      "inputTokenLimit": 0,
      "outputTokenLimit": 0,
      "supportedGenerationMethods": ["generateContent"]
    }
  ]
}

curl https://api.infillm.com/v1/moderations \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-moderation-latest",
    "input": "请检查这段文本是否合规。"
  }'

• 对用户生成内容、公开发布内容和工具执行前输入进行审查。
• 批量输入时，results 的顺序与 input 数组顺序一致。
• flagged: true 表示建议拦截、降级或进入人工复核流程。

wscat -c "wss://api.infillm.com/v1/realtime?model=gpt-4o-realtime-preview" \
  -H "Authorization: Bearer $INFILLM_API_KEY"

• Realtime 不是普通 HTTP JSON 接口，需要使用 WebSocket 客户端。
• 鉴权仍使用 API Key。
• 具体事件字段会随上游实时模型能力变化，接入时应保留对未知事件类型的兼容处理。

curl https://api.infillm.com/v1/rerank \
  -H "Authorization: Bearer $INFILLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "rerank-english-v2.0",
    "query": "什么是 API 网关？",
    "documents": [
      "API 网关统一处理鉴权、限流、路由和日志。",
      "向量数据库用于相似度搜索。",
      "负载均衡可以把请求分发到多个服务实例。"
    ],
    "top_n": 2,
    "return_documents": true
  }'

• 先用向量检索召回较多候选文档，再用 Rerank 精排。
• top_n 应按业务需要控制，避免对过多文档做不必要的精排。
• 如果应用已保存原文，可将 return_documents 设为 false 以减少响应体积。