gpt 上下文压缩

压缩 GPT 对话上下文，减少后续请求的 token 占用。

网站中的模型与调用示例

请求参数

{
  "contentType": "application/json",
  "properties": {
    "input": {
      "description": "需要压缩的历史消息列表",
      "items": {
        "properties": {
          "content": {
            "description": "消息内容",
            "type": "string",
            "x-i18n": {
              "en-US": {
                "description": "Message content"
              }
            }
          },
          "role": {
            "description": "消息角色，可选值：user / assistant / system",
            "type": "string",
            "x-i18n": {
              "en-US": {
                "description": "Message role, optional values: user/assistant/system"
              }
            }
          },
          "type": {
            "description": "消息类型，固定值为 message",
            "type": "string",
            "x-i18n": {
              "en-US": {
                "description": "Message type, fixed value is message"
              }
            }
          }
        },
        "required": [
          "type",
          "role",
          "content"
        ],
        "type": "object"
      },
      "type": "array",
      "x-i18n": {
        "en-US": {
          "description": "List of historical messages that need to be compressed"
        }
      }
    },
    "model": {
      "description": "模型名称，例如 gpt-5-codex",
      "type": "string",
      "x-i18n": {
        "en-US": {
          "description": "Model name, such as gpt-5-codex"
        }
      }
    }
  },
  "required": [
    "model",
    "input"
  ]
}

请求示例 · 示例 1

{
  "input": [
    {
      "content": [
        {
          "text": "Hello, this is a compact test.",
          "type": "input_text"
        }
      ],
      "role": "user"
    }
  ],
  "model": "gpt-5.4"
}

响应示例 · 成功示例

{
  "created_at": 1779183068,
  "id": "resp_010f1725ac1c2947016a0c2ddb9dcc81989aff546cbc9f7eb2",
  "object": "response.compaction",
  "output": [
    {
      "content": [
        {
          "text": "你好，请用一句话回复。",
          "type": "input_text"
        }
      ],
      "id": "msg_010f1725ac1c2947016a0c2ddbbd38819883f6c1d69c625183",
      "role": "user",
      "status": "completed",
      "type": "message"
    },
    {
      "encrypted_content": "gAAAAABqDC3c-W14MVMVz_8_5eeXDk1S-t1_rDOciRXR4HpuZXluWNcLMVLzW57ojCF4tl0-K5GdbKqIAQ1cBfr7FOY52gylXqqUE9uHbI3Pimil5CA0fhVGtiKZ_Me_VDUw8d_YU43C0qRBV4tzTeHszRVwNiod8v0HuqFE5B7XeDPtYcer_RIRO2qLbK0NvDlcx1LtiuHDyAjOKan3V-prukIJDx1-9oNTJmj6LX9aodXhHiltA1j49ufZuk7NsEzDkK0u588HDhraI0PCU-bY0Rpih7Yd2nE_XEnEfLpsYBzgyt1YoTObsOb1AIU0tBQ8OMw101dMXnv6EQ_xwp6mE1o7oHcho7AK86_MBZS4GehUef9nMZMJsQGGaGdjkC5eKED5XWVMAlQ5A5LlPrx3Ne3bScLUvQRNPZocAGrWbHUqGoGHoee6LkPhqpAkyV2JecvLynXR4kMhCQOaktUDbIC2Uszm2qkfWBkACH8yV_btuSPZhhk5KxrmCgi55JAA6esv1PK3kRrpj6Am6RkmqQH017koqsrUu62S8dtUlWn3hZY7kTSX9ISvL1gkoue05QPgM0TSeSLAPSax2MpqJUEajPQv0ktFKGZP88npcXYarb20Nw7k2IfchS79cQ90Y7-BLPLKErHHnssNUzdVgiCjzoCCQsSn28d52DGvtOJqfQ_JyHPBma8yKDpkRpLJggScKfyEpA-v6iTbRnQUY8TV29V37tZAKrMZ6p8dRWHYI0cehnfdqr0VPRFr-Z0FkX1rHM-0CKx-WyHQQ8_kdxh5WyaQq1a-J8R1SDQFMtU89wB6bdCBavi27GE63dKYfG6oJyetzNtnlJLBXkjAjTusSKCYdA==",
      "id": "cmp_010f1725ac1c2947016a0c2ddc48588198b65e0e0816f998a6",
      "type": "compaction_summary"
    }
  ],
  "usage": {
    "input_tokens": 31,
    "input_tokens_details": {
      "cached_tokens": 0
    },
    "output_tokens": 34,
    "output_tokens_details": {
      "reasoning_tokens": 0
    },
    "total_tokens": 65
  }
}

请求参数

Authorization

Bearer Token

在 Header 添加参数

Authorization

，其值为在 Bearer 之后拼接 Token

示例：

Authorization: Bearer ********************

Body 参数application/json必填

示例

gpt 上下文压缩gpt 上下文压缩 · 示例 1

{
    "model": "gpt-5.4",
    "input": [
        {
            "type": "message",
            "role": "user",
            "content": "Hello, this is a compact test."
        }
    ]
}

请求示例代码

Shell

JavaScript

Java

Swift

PHP

Python

HTTP

Objective-C

Ruby

OCaml

Dart

返回响应

🟢200成功

application/json

Bodyapplication/json

示例

{
    "id": "resp_010f1725ac1c2947016a0c2ddb9dcc81989aff546cbc9f7eb2",
    "object": "response.compaction",
    "created_at": 1779183068,
    "output": [
        {
            "id": "msg_010f1725ac1c2947016a0c2ddbbd38819883f6c1d69c625183",
            "type": "message",
            "status": "completed",
            "content": [
                {
                    "type": "input_text",
                    "text": "你好，请用一句话回复。"
                }
            ],
            "role": "user"
        },
        {
            "id": "cmp_010f1725ac1c2947016a0c2ddc48588198b65e0e0816f998a6",
            "type": "compaction_summary",
            "encrypted_content": "gAAAAABqDC3c-W14MVMVz_8_5eeXDk1S-t1_rDOciRXR4HpuZXluWNcLMVLzW57ojCF4tl0-K5GdbKqIAQ1cBfr7FOY52gylXqqUE9uHbI3Pimil5CA0fhVGtiKZ_Me_VDUw8d_YU43C0qRBV4tzTeHszRVwNiod8v0HuqFE5B7XeDPtYcer_RIRO2qLbK0NvDlcx1LtiuHDyAjOKan3V-prukIJDx1-9oNTJmj6LX9aodXhHiltA1j49ufZuk7NsEzDkK0u588HDhraI0PCU-bY0Rpih7Yd2nE_XEnEfLpsYBzgyt1YoTObsOb1AIU0tBQ8OMw101dMXnv6EQ_xwp6mE1o7oHcho7AK86_MBZS4GehUef9nMZMJsQGGaGdjkC5eKED5XWVMAlQ5A5LlPrx3Ne3bScLUvQRNPZocAGrWbHUqGoGHoee6LkPhqpAkyV2JecvLynXR4kMhCQOaktUDbIC2Uszm2qkfWBkACH8yV_btuSPZhhk5KxrmCgi55JAA6esv1PK3kRrpj6Am6RkmqQH017koqsrUu62S8dtUlWn3hZY7kTSX9ISvL1gkoue05QPgM0TSeSLAPSax2MpqJUEajPQv0ktFKGZP88npcXYarb20Nw7k2IfchS79cQ90Y7-BLPLKErHHnssNUzdVgiCjzoCCQsSn28d52DGvtOJqfQ_JyHPBma8yKDpkRpLJggScKfyEpA-v6iTbRnQUY8TV29V37tZAKrMZ6p8dRWHYI0cehnfdqr0VPRFr-Z0FkX1rHM-0CKx-WyHQQ8_kdxh5WyaQq1a-J8R1SDQFMtU89wB6bdCBavi27GE63dKYfG6oJyetzNtnlJLBXkjAjTusSKCYdA=="
        }
    ],
    "usage": {
        "input_tokens": 31,
        "input_tokens_details": {
            "cached_tokens": 0
        },
        "output_tokens": 34,
        "output_tokens_details": {
            "reasoning_tokens": 0
        },
        "total_tokens": 65
    }
}

🟠400请求有误

🟠401没有权限

gpt 上下文压缩

网站中的模型与调用示例#

gpt 上下文压缩#

请求参数#

请求示例 · 示例 1#

响应示例 · 成功示例#

请求参数

请求示例代码

返回响应

网站中的模型与调用示例

gpt 上下文压缩

请求参数

请求示例 · 示例 1

响应示例 · 成功示例