Gemini媒体识别

Gemini媒体识别

POST /v1beta/models/{model}:generateContent

Gemini图像、PDF、音频、视频识别请求

⚠️注意：仅支持通过 inlineData 以 base64 方式上传图像、PDF、音频、视频，不支持 fileData.fileUri 或 File API。

路径参数 (Path Params)

参数名	类型	必选	描述	示例
model	string	是	模型名称	``

请求头 (Headers)

参数名	类型	必选	描述	示例
Authorization	string	否	鉴权令牌	Bearer ********************

请求体 (Request Body)

Content-Type: application/json

字段	类型	必选	描述
contents	array of object	否
contents[].role	string	否
contents[].parts	array of object	否
contents[].parts[].text	string	否
contents[].parts[].inlineData	object	否
contents[].parts[].inlineData.mimeType	string	否
contents[].parts[].inlineData.data	string	否
generationConfig	object	否
generationConfig.temperature	number	否
generationConfig.topP	number	否
generationConfig.topK	integer	否
generationConfig.maxOutputTokens	integer	否
generationConfig.stopSequences	array of string	否
safetySettings	array of object	否
safetySettings[].category	string	否
safetySettings[].threshold	string	否
tools	array of object	否
systemInstruction	object	否
systemInstruction.parts	array of object	否

请求体示例

{
  "contents": [
    {
      "role": "string",
      "parts": [
        {
          "text": "string",
          "inlineData": {
            "mimeType": "string",
            "data": "string"
          }
        }
      ]
    }
  ],
  "generationConfig": {
    "temperature": 0,
    "topP": 0,
    "topK": 0,
    "maxOutputTokens": 0,
    "stopSequences": [
      "string"
    ]
  },
  "safetySettings": [
    {
      "category": "string",
      "threshold": "string"
    }
  ],
  "tools": [
    {}
  ],
  "systemInstruction": {
    "parts": [
      {}
    ]
  }
}

返回响应 (Responses)

200 - 成功

字段	类型	描述
candidates	array of object
candidates[].content	object
candidates[].content.role	string
candidates[].content.parts	array of object
candidates[].finishReason	string
candidates[].safetyRatings	array of object
usageMetadata	object
usageMetadata.promptTokenCount	integer
usageMetadata.candidatesTokenCount	integer
usageMetadata.totalTokenCount	integer

响应示例

{
  "candidates": [
    {
      "content": {
        "role": "string",
        "parts": [
          {}
        ]
      },
      "finishReason": "string",
      "safetyRatings": [
        {}
      ]
    }
  ],
  "usageMetadata": {
    "promptTokenCount": 0,
    "candidatesTokenCount": 0,
    "totalTokenCount": 0
  }
}