请求方式：POST
请求地址： https://{domain}/dataAgent/llm/openApi/v2/deepresearch/executeTask

[
    {"id": 1, "appId": 1},
    {"id": 2, "appId": 1}
]

[
    {
        "reportId": 1,
        "appId": 1,
        "dashboardId": 1,
        "sheetId": 1
    },
    {
        "reportId": 2,
        "appId": 1,
    }
]

{
    "schemes": ["Bearer"],
    "credentials": "sk-12345"
}

{
    "schemes": ["ApiKey"],
    "credentials": "key:value"
}

{
    "stream": true,
    "metadata": {
        "agentId": 17368,
        "reportList":[
            {
                "appId":1004033,
                "reportId": 7112515
            }
        ],
        "dataSetList": [
            {
                "id": 1836089,
                "appId": 1004033
            }
        ],
        "fileList": [
            "agent-upload-20250916121852-a79607aef5e44339ad9afa1aa41c81b5"
        ]
    },
    "content": "立即进入报告编写阶段，在报告编写阶段中通过render_to_chart的方法将传入的文件df渲染为图表，只需要这一个步骤即可，报告长度不超过100词"
}

非流式返回 #

{
    "JSONIFY_PRETTYPRINT_REGULAR": false,
    "code": "llm/ok",
    "data": {
        "taskId": "d9ec4b426c5a4f7d8e30ced79d5ecb0e"
    },
    "msg": {}
}

流式返回 #

data: {"task": {"kind": "task", "id": "314b788f83394219afa563e47a2938a9", "contextId": "fd86552d317b4733aa8e7438f81c84b5", "status": {"state": "submitted", "timestamp": "2025-09-26T14:50:42.280030"}, "metadata": {"agentId": 17368, "dataSetList": [{"id": 1836089, "appId": 1004033}], "reportList": [{"reportId": 7112515, "appId": 1004033}], "fileList": ["agent-upload-20250916121852-a79607aef5e44339ad9afa1aa41c81b5"]}}}

data: {"statusUpdate": {"kind": "status-update", "taskId": "314b788f83394219afa563e47a2938a9", "contextId": "fd86552d317b4733aa8e7438f81c84b5", "status": {"state": "working", "timestamp": "2025-09-26T14:50:45.351049"}, "final": false, "metadata": {"agentId": 17368, "sessionId": 12013496, "deepResearchTaskId": 181928, "dataSetList": [{"id": 1836089, "appId": 1004033}], "reportList": [{"reportId": 7112515, "appId": 1004033}], "fileList": ["agent-upload-20250916121852-a79607aef5e44339ad9afa1aa41c81b5"]}}}

data: {"statusUpdate": {"kind": "status-update", "taskId": "314b788f83394219afa563e47a2938a9", "contextId": "fd86552d317b4733aa8e7438f81c84b5", "status": {"state": "working", "timestamp": "2025-09-26T14:51:15.928361"}, "final": false, "metadata": {"agentId": 17368, "sessionId": 12013496, "deepResearchTaskId": 181928, "dataSetList": [{"id": 1836089, "appId": 1004033}], "reportList": [{"reportId": 7112515, "appId": 1004033}], "fileList": ["agent-upload-20250916121852-a79607aef5e44339ad9afa1aa41c81b5"]}}}

data: {"artifactUpdate": {"kind": "artifact-update", "taskId": "314b788f83394219afa563e47a2938a9", "contextId": "fd86552d317b4733aa8e7438f81c84b5", "artifact": {"artifactId": "821c905ef77c4e0ead2c22f4a20e4365", "name": "0x验证数据可视化报告", "description": "Markdown Report", "parts": [{"kind": "text", "text": "# 0x验证数据可视化报告\n\n# 1 数据概览与图表展示\n\n基于0x验证表的数据分析，该数据集包含3个日期（2023-04-20至2023-04-22），涵盖2个项目（A和B），共计6条记录。\n\n\n\n\n| 日期         | 项目   |    数字 |      浮点 |\n|------------|------|-------|---------|\n| 2023-04-20 | B    | 49930 | 39940.4 |\n| 2023-04-20 | A    |  9980 | 19980.5 |\n| 2023-04-21 | A    |  9990 | 19990.7 |\n| 2023-04-21 | B    | 49980 | 39980.6 |\n| 2023-04-22 | A    | 10000 | 20000.9 |\n| 2023-04-22 | B    | 50000 | 40000.8 |\n\n\n\n数据可视化显示，项目B的数字总和为149,910，显著高于项目A的29,970，两者比例约为5:1。浮点数据同样呈现类似分布趋势，项目B的浮点总和约为项目A的2倍。\n\n# 参考信息\n\n本报告的数据来源与图表引用如下："}], "metadata": {"type": "deep_research_markdown_report"}}}}

data: {"statusUpdate": {"kind": "status-update", "taskId": "314b788f83394219afa563e47a2938a9", "contextId": "fd86552d317b4733aa8e7438f81c84b5", "status": {"state": "completed", "timestamp": "2025-09-26T14:52:00.878760"}, "final": true, "metadata": {"agentId": 17368, "sessionId": 12013496, "deepResearchTaskId": 181928, "dataSetList": [{"id": 1836089, "appId": 1004033}], "reportList": [{"reportId": 7112515, "appId": 1004033}], "fileList": ["agent-upload-20250916121852-a79607aef5e44339ad9afa1aa41c81b5"]}}}

artifact定义 #

流式返回结果中，产物更新结果（artifactUpdate）中包含的artifact字段的内容示例及详细字段说明如下。

{
    "artifactId": "080e334f-29c1-5587-bd0c-e230bd78da0d",
    "name": "绘制各地区销售额分布图",
    "description": "任务已完成：【执行Python代码】绘制各地区销售额分布图",
    "parts": [
        {
            "kind": "text",
            "text": "......"
        },
        {
            "kind": "data",
            "data": {
                "type": "python_sandbox",
                "content": "```python......```\n\n",
                "result": "代码执行成功，输出内容如下：......"
            }
        }
    ],
    "metadata": {
        "type": "deep_research_running_step",
        "stepType": "python_sandbox",
        "stepId": -15913542,
        "parentStepId": 15913542
    }
}

中间步骤产物：deep_research_running_step #

executor_think #

• 不含data块
• 存在流式更新，更新至text块中

data: {"artifactUpdate": {"kind": "artifact-update", "taskId": "49e0fed9b5714bf9807164d012ce4b2a", "contextId": "0c4f27bcb95f40fd8a0e72c93647dca0", "artifact": {"artifactId": "9c73aa1e-4b6a-574a-92ae-5b3e7523e98a", "name": "思考推理", "description": "思考推理", "parts": [{"kind": "text", "text": ""}], "metadata": {"type": "deep_research_running_step", "stepType": "executor_think", "stepId": 15913627, "parentStepId": 15913542}}, "lastChunk": false}}

data: {"artifactUpdate": {"kind": "artifact-update", "taskId": "49e0fed9b5714bf9807164d012ce4b2a", "contextId": "0c4f27bcb95f40fd8a0e72c93647dca0", "artifact": {"artifactId": "9c73aa1e-4b6a-574a-92ae-5b3e7523e98a", "parts": [{"kind": "text", "text": "##"}]}, "append": true, "lastChunk": false}}

......

data: {"artifactUpdate": {"kind": "artifact-update", "taskId": "49e0fed9b5714bf9807164d012ce4b2a", "contextId": "0c4f27bcb95f40fd8a0e72c93647dca0", "artifact": {"artifactId": "9c73aa1e-4b6a-574a-92ae-5b3e7523e98a", "parts": [{"kind": "text", "text": "。"}]}, "append": true, "lastChunk": false}}

data: {"artifactUpdate": {"kind": "artifact-update", "taskId": "49e0fed9b5714bf9807164d012ce4b2a", "contextId": "0c4f27bcb95f40fd8a0e72c93647dca0", "artifact": {"artifactId": "9c73aa1e-4b6a-574a-92ae-5b3e7523e98a", "name": "思考推理", "description": "思考推理完成", "parts": [{"kind": "text", "text": "## 思路分析\n\n已完成各地区销售额的查询和分析，获得了61个城市的数据，并绘制了各省份销售额分布图。数据显示：\n- 广东省和江苏省销售额领先\n- 前10名省份占总销售额74.37%\n- 已通过可视化展示分布情况\n\n任务已完成，可以进入报告编写阶段。"}], "metadata": {"type": "deep_research_running_step", "stepType": "executor_think", "stepId": 15913627, "parentStepId": 15913542}}, "lastChunk": true}}

table_insight #

• 不含data块

data: {"artifactUpdate": {"kind": "artifact-update", "taskId": "49e0fed9b5714bf9807164d012ce4b2a", "contextId": "0c4f27bcb95f40fd8a0e72c93647dca0", "artifact": {"artifactId": "86875483-3508-56b3-8f16-7cb041226e2a", "name": "获取数据集相关信息", "description": "任务已完成：获取数据集相关信息", "parts": [{"kind": "text", "text": "\n\n## 表`1836089`的相关信息\n表信息如下:\n表ID: `1836089`\n表名: `销售GMV-直连`\n\n\n引擎列名和数据类型：\n|列名|数据类型|\n| ---- | ---- |\n|sale_price|Nullable(Float32)|\n|order_count|int|\n\n\n基于引擎列创建的字段列表，可以直接使用，字段名均以``包围：\n\n|字段名|数据类型|相关信息|抽样示例值(非全量)|\n| ---- | ---- | ---- | ---- |\n|`销售年份`|int|||\n|`customer_id`|int|||\n|`order_date`|date|数据可用日期范围：最大值: 2016-06-30,最小值: 2013-07-01||\n|`sale_price`|float|||\n|`order_count`|int|||\n|`price`|float|||\n|`null维度`|string|||\n|`product_id`|string|||\n|`图表中没有出现的公共维度`|string|||\n|`sale_city`|string|||\n|`公共维度`|string|||\n|`产品类型`|string|||\n|`省份`|string|||\n|`GMV`|float|||\n\n\n聚合指标定义参考:\n\n```\n`总销售额`: sum(((order_count) * (sale_price)))\n```\n\n\n## 表`2133291`的相关信息\n表信息如下:\n表ID: `2133291`\n表名: `0x验证表`\n\n字段列表，字段名均以``包围：\n\n|字段名|数据类型|相关信息|抽样示例值(非全量)|\n| ---- | ---- | ---- | ---- |\n|`日期`|date|数据可用日期范围：最大值: 2023-04-22,最小值: 2023-04-20||\n|`数字`|int|||\n|`浮点`|float|||\n\n\n\n"}], "metadata": {"type": "deep_research_running_step", "stepType": "table_insight", "stepId": -15913420, "parentStepId": 15913420}}, "lastChunk": true}}

sql_sandbox #

• data块定义

data: {"artifactUpdate": {"kind": "artifact-update", "taskId": "49e0fed9b5714bf9807164d012ce4b2a", "contextId": "0c4f27bcb95f40fd8a0e72c93647dca0", "artifact": {"artifactId": "81301578-a38a-5df9-aeb6-403249f7b650", "name": "查询各个地区的销售额数据", "description": "任务已完成：【执行SQL语句】按地区维度查询销售额数据，包括省份和城市信息，计算总销售额", "parts": [{"kind": "text", "text": "```sql\nSELECT `省份`,\n       `sale_city`,\n       SUM(`order_count` * `sale_price`) as `总销售额`\nFROM   `销售GMV-直连`\nWHERE  `省份` IS NOT NULL\n   AND `sale_city` IS NOT NULL\nGROUP BY `省份`, `sale_city`\nORDER BY `总销售额` DESC\n```\n\n查询结果[sales_by_region]共61行，详情如下：\n||省份|sale_city|总销售额|\n|--|--|--|--|\n|0|上海市|上海市|2832629.8676199913|...查询成功。查询结果已存储在内存变量sales_by_region中，类型为pandas.DataFrame\nsales_by_region的列名和列数据类型如下:\n|列名|数据类型|\n|--|--|\n|省份|string|\n|sale_city|string|\n|总销售额|float|\n\n查询结果sales_by_region的总行数为：61\n\nsales_by_region的完整的结果数据如下：\n||省份|sale_city|总销售额|\n|--|--|--|--|\n|0|上海市|上海市|2832629.8676199913|..."}, {"kind": "data", "data": {"type": "execute_sql", "content": "```sql\nSELECT `省份`,\n       `sale_city`,\n       SUM(`order_count` * `sale_price`) as `总销售额`\nFROM   `销售GMV-直连`\nWHERE  `省份` IS NOT NULL\n   AND `sale_city` IS NOT NULL\nGROUP BY `省份`, `sale_city`\nORDER BY `总销售额` DESC\n```\n\n", "result": "查询结果[sales_by_region]共61行，详情如下：\n||省份|sale_city|总销售额|\n|--|--|--|--|\n|0|上海市|上海市|2832629.8676199913|", "executeSql": "SELECT (sale_province) AS `alias_prefix_省份`,\n       (sale_city) AS `alias_prefix_sale_city`,\n       SUM((order_count) * (sale_price)) AS `alias_prefix_总销售额`\nFROM   (SELECT `product_type` as `product_type`,\n               `product_id` as `product_id`,\n               `order_count` as `order_count`,\n               `price` as `price`,\n               `sale_province` as `sale_province`,\n               `sale_price` as `sale_price`,\n               `sale_city` as `sale_city`,\n               `order_date` as `order_date`,\n               `customer_id` as `customer_id`\n        FROM   `aeolus_excel_upload_aeolus_rei`.`table_cede732c_c35a_424c_85e7_174fee1739ae`\n        WHERE  1 = 1) AS `销售GMV-直连`\nWHERE  NOT ((sale_province) IS NULL)\n   AND NOT ((sale_city) IS NULL)\nGROUP BY `alias_prefix_省份`, `alias_prefix_sale_city`\nORDER BY `alias_prefix_总销售额` DESC LIMIT 10000 FORMAT JSONCompact", "rewriteSql": "SELECT (`1700031515177_省份`) AS `alias_prefix_省份`,\n       (`1700031515178_sale_city`) AS `alias_prefix_sale_city`,\n       SUM((`1700031515180_order_count`) * (`1700031515182_sale_price`)) AS `alias_prefix_总销售额`\nFROM   `chenrui.ovo 测试.销售GMV-直连` AS `销售GMV-直连`\nWHERE  NOT ((`1700031515177_省份`) IS NULL)\n   AND NOT ((`1700031515178_sale_city`) IS NULL)\nGROUP BY `alias_prefix_省份`, `alias_prefix_sale_city`\nORDER BY `alias_prefix_总销售额` DESC LIMIT 10000"}}], "metadata": {"type": "deep_research_running_step", "stepType": "sql_sandbox", "stepId": -15913462, "parentStepId": 15913462}}, "lastChunk": true}}

python_sandbox #

• data块定义如下

data: {"artifactUpdate": {"kind": "artifact-update", "taskId": "49e0fed9b5714bf9807164d012ce4b2a", "contextId": "0c4f27bcb95f40fd8a0e72c93647dca0", "artifact": {"artifactId": "080e334f-29c1-5587-bd0c-e230bd78da0d", "name": "绘制各地区销售额分布图", "description": "任务执行中：【执行Python代码】绘制各地区销售额分布图", "parts": [{"kind": "text", "text": ""}, {"kind": "data", "data": {"type": "python_sandbox", "content": "```python\nimport matplotlib.pyplot as plt\nimport pandas as pd\n\n# 设置中文字体\nplt.rcParams['font.sans-serif'] = ['SimHei']\nplt.rcParams['axes.unicode_minus'] = False\n\n# 按省份汇总销售额\nprovince_sales = sales_by_region.groupby('省份')['总销售额'].sum().sort_values(ascending=False)\n\n# 绘制柱状图\nplt.figure(figsize=(12, 8))\nprovince_sales.plot(kind='bar', color='skyblue')\nplt.title('各省份销售额分布')\nplt.xlabel('省份')\nplt.ylabel('销售额')\nplt.xticks(rotation=45)\nplt.tight_layout()\nplt.show()\n\n# 输出前10名省份\nprint(\"销售额前10名省份:\")\nprint(province_sales.head(10))\n\n# 计算前10名省份占总销售额的比例\ntop10_ratio = province_sales.head(10).sum() / province_sales.sum()\nprint(f\"\\n前10名省份占总销售额比例: {top10_ratio:.2%}\")\n```\n\n"}}], "metadata": {"type": "deep_research_running_step", "stepType": "python_sandbox", "stepId": -15913542, "parentStepId": 15913542}}, "lastChunk": false}}

data: {"artifactUpdate": {"kind": "artifact-update", "taskId": "49e0fed9b5714bf9807164d012ce4b2a", "contextId": "0c4f27bcb95f40fd8a0e72c93647dca0", "artifact": {"artifactId": "080e334f-29c1-5587-bd0c-e230bd78da0d", "name": "绘制各地区销售额分布图", "description": "任务已完成：【执行Python代码】绘制各地区销售额分布图", "parts": [{"kind": "text", "text": "```python\nimport matplotlib.pyplot as plt\nimport pandas as pd\n\n# 设置中文字体\nplt.rcParams['font.sans-serif'] = ['SimHei']\nplt.rcParams['axes.unicode_minus'] = False\n\n# 按省份汇总销售额\nprovince_sales = sales_by_region.groupby('省份')['总销售额'].sum().sort_values(ascending=False)\n\n# 绘制柱状图\nplt.figure(figsize=(12, 8))\nprovince_sales.plot(kind='bar', color='skyblue')\nplt.title('各省份销售额分布')\nplt.xlabel('省份')\nplt.ylabel('销售额')\nplt.xticks(rotation=45)\nplt.tight_layout()\nplt.show()\n\n# 输出前10名省份\nprint(\"销售额前10名省份:\")\nprint(province_sales.head(10))\n\n# 计算前10名省份占总销售额的比例\ntop10_ratio = province_sales.head(10).sum() / province_sales.sum()\nprint(f\"\\n前10名省份占总销售额比例: {top10_ratio:.2%}\")\n```\n\n代码执行成功，输出内容如下：\n\n销售额前10名省份:\n                  总销售额\n省份                    \n广东省  4822876.405101776\n江苏省  4590771.994219542\n上海市 2832629.8676199913\n北京市 2738369.6146588326\n浙江省 2471835.6385302544\n河北省  1499687.414835453\n福建省  936069.9314939976\n天津市  687273.1442761421\n陕西省  615094.7341973782\n重庆市  602230.6761865616\n\n前10名省份占总销售额比例: 74.37%\n\n已渲染图片：\n  - title：各省份销售额分布\n  - img_key： b2134d0c161e4822a7d697271802a1d4\n  - img_url：https://data.bytedance.net/dataAgent/llm/api/v2/deepresearch/images?img_key=13464321-286833-b2134d0c161e4822a7d697271802a1d4\n\n"}, {"kind": "data", "data": {"type": "python_sandbox", "content": "```python\nimport matplotlib.pyplot as plt\nimport pandas as pd\n\n# 设置中文字体\nplt.rcParams['font.sans-serif'] = ['SimHei']\nplt.rcParams['axes.unicode_minus'] = False\n\n# 按省份汇总销售额\nprovince_sales = sales_by_region.groupby('省份')['总销售额'].sum().sort_values(ascending=False)\n\n# 绘制柱状图\nplt.figure(figsize=(12, 8))\nprovince_sales.plot(kind='bar', color='skyblue')\nplt.title('各省份销售额分布')\nplt.xlabel('省份')\nplt.ylabel('销售额')\nplt.xticks(rotation=45)\nplt.tight_layout()\nplt.show()\n\n# 输出前10名省份\nprint(\"销售额前10名省份:\")\nprint(province_sales.head(10))\n\n# 计算前10名省份占总销售额的比例\ntop10_ratio = province_sales.head(10).sum() / province_sales.sum()\nprint(f\"\\n前10名省份占总销售额比例: {top10_ratio:.2%}\")\n```\n\n", "result": "代码执行成功，输出内容如下：\n\n销售额前10名省份:\n\n\n\n| 省份   | 总销售额               |\n|:-----|:-------------------|\n| 广东省  | 4822876.405101776  |\n| 江苏省  | 4590771.994219542  |\n| 上海市  | 2832629.8676199913 |\n| 北京市  | 2738369.6146588326 |\n| 浙江省  | 2471835.6385302544 |\n| 河北省  | 1499687.414835453  |\n| 福建省  | 936069.9314939976  |\n| 天津市  | 687273.1442761421  |\n| 陕西省  | 615094.7341973782  |\n| 重庆市  | 602230.6761865616  |\n\n\n\n\n前10名省份占总销售额比例: 74.37%\n\n\n\n\n\n![已渲染图片\"各省份销售额分布\"，可通过地址 https://data.bytedance.net/dataAgent/llm/api/v2/deepresearch/images?img_key=13464321-286833-b2134d0c161e4822a7d697271802a1d4 访问](https://data.bytedance.net/dataAgent/llm/api/v2/deepresearch/images?img_key=13464321-286833-b2134d0c161e4822a7d697271802a1d4 \"各省份销售额分布\")\n\n\n\n\n\n"}}], "metadata": {"type": "deep_research_running_step", "stepType": "python_sandbox", "stepId": -15913542, "parentStepId": 15913542}}, "lastChunk": true}}

拼接深度研究报告URL #

以下为您介绍执行深度研究任务openAPI后，如何基于返回结果来拼接最终生成的研究报告URL，拼接得到报告URL后即可通过此URL展示网页版报告内容。
首先我们需要了解，当前拼接深度研究报告URL时，有以下两种格式。

拼接深度研究报告URL时，建议您先打开一个智能分析Agent的会话页面，复制页面URL，在此基础之上进行修改拼接，能快速完成URL的拼接。以下为您示例操作步骤。

1. 登录并进入智能分析Agent的某个会话页面，比如，打开智能问数的问题提问页面，然后复制浏览器中的页面URL。
   例如，复制的页面URL为：https://console.volcengine.com/bi/datawind/data-agent/analytics-agent/pages/agent/session?agentId=2730&appId=1000326&forceSwitchToShowAgent=true&region=cn-shanghai&sessionId=258880。

2. 对复制得到的URL进行修改，得到拼接后的深度研究报告URL。

   对比说明	方式1：SSO登录场景	方式2：单独认证场景
   最终报告URL	https://console.volcengine.com/bi/datawind/data-agent/analytics-agent/pages/agent/report/html?region=cn-shanghai&reportKey=XXX	https://console.volcengine.com/bi/datawind/data-agent/analytics-agent-open/pages/agent/report/html?region=cn-shanghai&reportKey=XXX&jwtToken=YYY
   修改说明	将获取到的会话页面URL的session？开始，往后的内容均删除，拼接报告的report/html?region=cn-shanghai&reportKey=XXX部分。 拼接时各变量参数的说明及获取方式见上文，其中： reportKey=XXX的取值为执行深度研究报告OpenAPI返回结果中的深度研究报告ID。 region为当前使用的智能体所在的地域信息。 URL的其余部分，其中，获取到的会话页面URL中的https://……bi/datawind部分为域名{domain}的取值，您无需修改，以页面复制结果为准。 URL的其余部分，其中，获取到的会话页面URL中{domain}与agent中间部分即为路径参数（data-agent/……/pages部分），您无需修改，以页面复制结果为准。	将获取到的会话页面URL的session？开始，往后的内容均删除，拼接报告的report/html?region=cn-shanghai&reportKey=XXX&jwtToken=YYY部分。 拼接时各变量参数的说明及获取方式见上文，其中： reportKey=XXX的取值为执行深度研究报告OpenAPI返回结果中的深度研究报告ID。 jwtToken=YYY为调用OpenAPI的jwttokens。 region为当前使用的智能体所在的地域信息。 URL的其余部分，其中，获取到的会话页面URL中的https://……bi/datawind部分为域名{domain}的取值，您无需修改，以页面复制结果为准。 URL的其余部分，其中，获取到的会话页面URL中{domain}与agent中间部分即为路径参数（data-agent/……//pages/部分），您需在复制得到的路径参数的analytics-agent后加上-open，作为此访问方式下的最终路径参数。