{"componentChunkName":"component---src-templates-acg-portal-new-template-tsx","path":"/nmwbl4w38","result":{"data":{"markdownRemark":{"html":"\n
\n
\n
\n
\n
POST
\n
\nhttps://qianfan.baidubce.com/v2/agent/deepresearch/run\n
\n
\n
\n

本接口用于与“深度研究”智能体进行多轮流式对话。“深度研究”是一个涉及多步推理与执行的长流程任务，不同于普通的“一问一答”。其完整的执行生命周期包含以下环节：

\n
    \n
  1. 发起研究：用户输入研究问题，支持上传背景文件（可选）。
    2. \n
  2. 需求澄清（非必经流程）：若用户请求研究的问题模糊、不完整或存在歧义，Agent多次主动与用户交互，通过提问来明确研究方向与方法。
    3. \n
  3. 大纲生成与确认：Agent生成研究大纲，用户可修改或确认。
    4. \n
  4. 深度执行：Agent将复杂问题逐步拆解，再结合联网信息及上传文件的混合检索拆解解决问题。
    5. \n
  5. 报告生成：输出长文本研究报告。
    6. \n
\n
\n
\n

权限说明

\n

调用本文API，需使用API Key鉴权方式。使用API Key鉴权调用API流程，具体调用流程，请查看认证鉴权。

\n
\n
\n

请求参数

\n
\n
\n
\n
\n
Headers 参数
\n
除公共头域外，无其它特殊头域
\n
\n
\n
\n
\n
Body 参数
\n
\n
\n
\n
\n
\n\nquery\n\nstring\n
\n

输入的对话信息。根据用户输入的query判断是简单问答还是进行深度研究。
\n由路由模块自动决定下一步是进入“需求澄清”阶段还是直接进入“大纲生成与确认”阶段。

\n
\n
必选
\n
\n
\n
\n\nagent_id\n\nstring\n
\n

请到千帆控制台Agent开发-官方Agent-深度研究Agent中，创建Agent并完成配置及发布后，回到我的Agent页面，即可复制ID

\n
\n
可选
\n
\n
\n
\n\nversion\n\nstring\n
\n

Agent版本

\n
    \n
  • lite：轻量版，默认版本
    • \n
  • standard：标准版
    • \n
  • pro：高级版
    • \n
\n
\n
可选
\n
\n
\n
\n\nfile_ids\n\narray\n
\n

文件ID列表。
\n若本轮对话需传入文件，需先通过对话创建接口获取conversation_id，再通过文件上传接口获取文件ID列表。
\n通过文件解析提交接口和文件解析查询接口可获取文件解析状态。

\n
\n
可选
\n
\n\n
\n\n显示子属性\n隐藏子属性\n
\n\n
\n
\n
\n\nitems\n\nstring\n
\n
\n
\n
\n
\n
\n
\n
\n\ninterrupt_id\n\nstring\n
\n

中断/干预 ID。
\n在回复 Agent 的\"需求澄清\"或\"大纲确认\"请求时必填。该字段内容在上一轮 SSE 响应中 content.text.data字段内获取。

\n
\n
可选
\n
\n
\n
\n\nconversation_id\n\nstring\n
\n

对话ID。
\n首轮对话：可不填，由服务端自动生成并在响应中返回，也可通过对话创建接口预先获取对话ID。
\n后续对话：必须填写，以保持上下文连续性，可在sse数据流获取conversation_id。

\n
\n
可选
\n
\n
\n
\n\nstructured_outline\n\nobject {4}\n
\n

深度研究报告的结构化大纲。
\n若上一轮对话输出为大纲确认数据，用户可基于大纲内容进行修改，下一轮对话请求携带最终大纲即可。Agent 将严格依据此数据生成后续的研究报告。

\n
\n
可选
\n
\n\n
\n\n显示子属性\n隐藏子属性\n
\n\n
\n
\n
\n\ntitle\n\nstring\n
\n

报告标题，可修改此字段以调整报告主题。

\n
\n
必选
\n
\n
\n
\n\nlocale\n\nstring\n
\n

指定报告生成的目标语言，目前仅支持简体中文。

\n
\n
可选
\n
\n
\n
\n\ndescription\n\nstring\n
\n

报告摘要，指导 Agent 全文写作。

\n
\n
必选
\n
\n
\n
\n\nsub_chapters\n\narray\n
\n

子章节列表，支持递归嵌套，章节当前支持2级展开。

\n
\n
必选
\n
\n\n
\n\n显示子属性\n隐藏子属性\n
\n\n
\n
\n
\n\nitems\n\nobject {3}\n
\n
\n
\n\n
\n\n显示子属性\n隐藏子属性\n
\n\n
\n
\n
\n\ntitle\n\nstring\n
\n

章节标题。

\n
\n
必选
\n
\n
\n
\n\ndescription\n\nstring\n
\n

章节写作指引，可对章节内容、组织结构、语言风格等给出明确要求。

\n
\n
必选
\n
\n
\n
\n\nsub_chapters\n\narray\n
\n

子章节列表，支持递归嵌套。

\n
\n
必选
\n
\n\n
\n\n显示子属性\n隐藏子属性\n
\n\n
\n
\n
\n\nitems\n\nobject {0}\n
\n
\n
\n\n
\n\n显示子属性\n隐藏子属性\n
\n\n暂无参数\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
请求结构
\n\n
\n
\n\n
\n
POST /v2/agent/deepresearch/run HTTP/1.1\nHOST: qianfan.baidubce.com\nAuthorization: Bearer <API Key>\nContent-Type: application/json\n{\n    \"query\": \"研究AI市场\",\n    \"version\": \"lite\"\n}
\n
\n
\n
\n
\n
\n
\n
\n

示例代码

\n
\n
\n\n\n\n
\n
\n
\n
\n\n
\n
\n
\n
Curl
\n\n
\n
\n\n
\n
curl -i --location 'https://qianfan.baidubce.com/v2/agent/deepresearch/run' \\\n--header 'Authorization: Bearer bce-v3/ALTAK-********/86356c2efcefb5776024ddb7d7958640f2fc16bf' \\\n--header 'Content-Type: application/json' \\\n--data '{\n    \"query\": \"研究AI市场\",\n    \"version\": \"lite\"\n}'
\n
\n
\n
\n
\n\n
\n
\n
\n
Curl
\n\n
\n
\n\n
\n
curl --location --request POST 'https://qianfan.baidubce.com/v2/agent/deepresearch/run' \\\n--header 'Authorization: Bearer bce-v3/ALTAK-********/86356c2efcefb5776024ddb7d7958640f2fc16bf' \\\n--header 'Content-Type: application/json' \\\n--data '{\n    \"query\": \"跳过\", \n    \"version\": \"lite\",\n    \"conversation_id\": \"5f54660a-****-****-bc99-e0c0d6db8a7d\",\n    \"interrupt_id\": \"83e03dfa-****-****-80cb-1ee991c219ae\"\n}'
\n
\n
\n
\n
\n\n
\n
\n
\n
Curl
\n\n
\n
\n\n
\n
curl --location --request POST 'https://qianfan.baidubce.com/v2/agent/deepresearch/run' \\\n--header 'Authorization: Bearer bce-v3/ALTAK-********/86356c2efcefb5776024ddb7d7958640f2fc16bf' \\\n--header 'Content-Type: application/json' \\\n--data '{\n    \"query\": \"确认\",\n    \"version\": \"lite\",\n    \"conversation_id\": \"5f54660a-****-****-bc99-e0c0d6db8a7d\",\n    \"interrupt_id\": \"83e03dfa-****-****-80cb-1ee991c219ae\",\n    \"structured_outline\": {\n        \"locale\": \"简体中文\",\n        \"title\": \"全球人工智能市场研究与战略分析报告\",\n        \"description\": \"本报告全面分析全球AI市场的发展现状、技术趋势、应用场景、竞争格局，并提供面向企业决策者、投资者和行业研究者的未来3-5年预测与投资建议。\",\n        \"sub_chapters\": [\n            {\n                \"title\": \"全球AI市场概述与技术趋势\",\n                \"description\": \"回顾过去5年AI市场发展，分析当前全球市场规模、增长驱动因素及核心技术趋势，包括大模型和多模态AI的演进。\",\n                \"sub_chapters\": [\n                    {\n                        \"title\": \"市场发展回顾与现状\",\n                        \"description\": \"总结过去5年全球AI市场的关键里程碑、增长数据和主要推动力。\",\n                        \"sub_chapters\": []\n                    },\n                    {\n                        \"title\": \"核心技术趋势分析\",\n                        \"description\": \"深入探讨大模型、多模态AI及智能体技术的最新进展和未来方向。\",\n                        \"sub_chapters\": []\n                    },\n                    {\n                        \"title\": \"增长驱动因素与挑战\",\n                        \"description\": \"识别市场增长的主要动力（如投资、政策）和潜在风险（如技术瓶颈、伦理问题）。\",\n                        \"sub_chapters\": []\n                    }\n                ]\n            }\n        ]\n    }\n}'
\n
\n
\n
\n
\n
\n
\n
\n

返回响应

\n
\n
\n
\n
\n
Headers 参数
\n
除公共头域外，无其它特殊头域
\n
\n
\n
\n
\n
返回参数
\n
\n
\n
\n
\n
\n\n响应参数\n\n\n
\n

本接口响应数据参见通用对话响应参数接口文档。响应内容为 SSE 流式数据，每一行 (data:) 为一个 JSON 对象。
\n
\n在Agent执行过程中，通过content.event.name字段来判断当前的执行阶段：
\n

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
阶段关键事件标识逻辑说明
需求澄清/toolcall/interruptAgent 挂起任务，主动向用户提问以收集关于研究范围的反馈。
大纲确认/toolcall/structured_outlineAgent 生成初步报告大纲，并进入等待状态，等待用户对大纲的反馈与确认。
\n
\n
可选
\n
\n
\n
\n
\n
\n
\n
\n
\n\n\n\n
\n
\n
\n
\n\n
\n
\n
\n
JSON
\n\n
\n
\n\n
\n
data: {\"role\": \"assistant\", \"status\": \"running\", \"content\": [{\"name\": \"/chat/chat_agent\", \"type\": \"text\", \"text\": {\"info\": \"您希望这份AI市场调研报告侧重于 **技术发展趋势**（如大模型、多模态），还是 **商业应用场景**（如金融、医疗）？\"}, \"event\": {\"id\": \"msg-01\", \"status\": \"running\", \"name\": \"/chat/chat_agent\"}}]}\n\ndata: {\"role\": \"assistant\", \"status\": \"running\", \"content\": [{\"name\": \"/chat/chat_agent\", \"type\": \"text\", \"text\": {\"info\": \"\\n\\n另外，您对报告的输出形式有特殊偏好吗？比如侧重数据图表，还是深度案例分析？\"}, \"event\": {\"id\": \"msg-01\", \"status\": \"running\", \"name\": \"/chat/chat_agent\"}}]}\n\ndata: {\"role\": \"assistant\", \"status\": \"running\", \"content\": [{\"name\": \"/chat/chat_agent\", \"type\": \"text\", \"text\": {\"info\": \"\\n\\n您也可以提出其他具体方向。如果您希望直接进入下一步，请输入“跳过”~\"}, \"event\": {\"id\": \"msg-01\", \"status\": \"running\", \"name\": \"/chat/chat_agent\"}}]}\n\ndata: {\"role\": \"assistant\", \"status\": \"done\", \"content\": [{\"name\": \"/chat/chat_agent\", \"type\": \"text\", \"text\": {\"info\": \"\"}, \"event\": {\"id\": \"msg-01\", \"status\": \"done\", \"name\": \"/chat/chat_agent\"}}]}\n\ndata: {\"role\": \"tool\", \"status\": \"interrupt\", \"content\": [{\"name\": \"/toolcall/interrupt\", \"type\": \"json\", \"text\": {\"data\": \"{\\\"interrupt_id\\\":\\\"1b6696fd-1f68-4a3e-a3e3-a6d771504f08\\\",\\\"interrupt_type\\\":\\\"deep_research_coordinator\\\",\\\"wait_for\\\":[\\\"query\\\"]}\"}, \"event\": {\"id\": \"evt-int-01\", \"status\": \"interrupt\", \"name\": \"/toolcall/interrupt\"}}]}
\n
\n
\n
\n
\n\n
\n
\n
\n
JSON
\n\n
\n
\n\n
\n
data: {\"role\": \"tool\", \"status\": \"done\", \"content\": [{\"type\": \"json\", \"name\": \"\", \"text\": {\"data\": \"{\\\"thinking\\\":\\\"研究主题是AI市场调研报告，核心需求是让我自行选择合适的调研维度来生成报告...\\\",\\\"locale\\\":\\\"简体中文\\\",\\\"title\\\":\\\"2025-2026年人工智能市场综合调研与趋势前瞻报告\\\",\\\"description\\\":\\\"本报告系统分析全球AI市场的规模结构、产业链生态...\\\",\\\"sub_chapters\\\":[{\\\"title\\\":\\\"全球AI市场总体规模与增长态势\\\",\\\"description\\\":\\\"综合分析全球及中国AI市场的整体规模...\\\",\\\"sub_chapters\\\":[...]}]}\"}, \"event\": {\"id\": \"f316fa4a-c6ec-4f88-97ae-2554afc78715\", \"status\": \"done\", \"name\": \"/toolcall/structured_outline\"}}]}\n\ndata: {\"role\": \"tool\", \"status\": \"interrupt\", \"content\": [{\"type\": \"json\", \"name\": \"\", \"text\": {\"data\": \"{\\\"interrupt_id\\\":\\\"1b6696fd-1f68-4a3e-a3e3-a6d771504f08\\\",\\\"interrupt_type\\\":\\\"deep_research_outliner\\\",\\\"wait_for\\\":[\\\"structured_outline\\\",\\\"query\\\"]}\"}, \"event\": {\"id\": \"6c57deb8-877c-4942-8e70-7cff37fbea88\", \"status\": \"interrupt\", \"name\": \"/toolcall/interrupt\"}}]}\n
\n
\n
\n
\n
\n\n
\n
\n
\n
JSON
\n\n
\n
\n\n
\n
data: {\"role\": \"tool\", \"status\": \"running\", \"content\": [{\"name\": \"/thought/plan\", \"type\": \"text\", \"text\": {\"info\": \"正在进行任务【AI市场调研报告】的规划。\"}, \"event\": {\"id\": \"d1d4ac0e-6153-43bb-8715-b9e8ce7f9029\", \"status\": \"running\", \"name\": \"/thought/plan\"}}]}\n\ndata: {\"role\": \"tool\", \"status\": \"running\", \"content\": [{\"name\": \"/thought/plan\", \"type\": \"plan\", \"text\": {\"tasks\": [{\"tool\": \"DeepSearchAgent\", \"id\": \"2214dbe5-d493-436c-a9fc-cc444b9076e1\", \"name\": \"全球AI市场规模、历史数据与驱动因素研究\", \"status\": \"preparing\"}, {\"name\": \"AI核心技术趋势与竞争格局深度分析\", \"status\": \"preparing\", \"tool\": \"DeepSearchAgent\", \"id\": \"40956565-f344-4bf0-ab94-6f4579388204\"}, {\"status\": \"preparing\", \"tool\": \"DeepSearchAgent\", \"id\": \"5f583630-ac62-4b0a-890b-660e2c3312e6\", \"name\": \"AI市场未来预测与战略建议信息收集\"}]}, \"event\": {\"id\": \"d1d4ac0e-6153-43bb-8715-b9e8ce7f9029\", \"status\": \"running\", \"name\": \"/thought/plan\"}}]}\n\ndata: {\"role\": \"tool\", \"status\": \"running\", \"content\": [{\"name\": \"/deep_search_agent\", \"type\": \"text\", \"text\": {\"info\": \"DeepSearch Agent1: 全球AI市场规模、历史数据与驱动因素研究\"}, \"event\": {\"id\": \"1f23e503-ce73-4e35-914a-babdf111ffc6\", \"status\": \"running\", \"name\": \"/deep_search_agent\"}}]}\n\ndata: {\"role\": \"tool\", \"status\": \"done\", \"content\": [{\"name\": \"/toolcall/baidu_search\", \"type\": \"references\", \"text\": {\"title\": \"IDC:全球人工智能总投资规模有望在2029年增长至12619亿美元\", \"content\": \"数据显示，2024年全球人工智能总投资规模为3159亿美元，\"}, \"event\": {\"id\": \"270b4ed6-174e-4a66-8601-0ab53a550d72\", \"status\": \"done\", \"name\": \"/toolcall/baidu_search\"}}]}\n\ndata: {\"role\": \"tool\", \"status\": \"done\", \"content\": [{\"name\": \"/toolcall/baidu_search\", \"type\": \"references\", \"text\": {\"title\": \"IDC:2024年全球人工智能IT总投资规模为3158亿美元 有望在2028年增至8159亿美元\", \"content\": \" IDC数据显示,2024年全球人工智能(AI)IT总投资规模为3,158亿美元,并有望在2028年增至8,159亿美元,五年复合增长率(CAGR)为32.9%。\"}, \"event\": {\"id\": \"b764b187-720a-4ce5-b843-8a6ef41129ae\", \"status\": \"done\", \"name\": \"/toolcall/baidu_search\"}}]}\n\ndata: {\"role\": \"tool\", \"status\": \"done\", \"content\": [{\"name\": \"/toolcall/baidu_search\", \"type\": \"references\", \"text\": {\"title\": \"斯坦福 AI 指数报告发布：2024年美国 AI 私人投资约为中国的 12 倍、GPT3.5 级别模型推理成本下降 280 倍\", \"content\": \"2024年全球私人AI投资达2523亿美元，其中生成式AI吸引339亿美元，同比增长18.7%。\"}, \"event\": {\"id\": \"97152d46-f8bb-4ce9-86ef-4812937bc8ca\", \"status\": \"done\", \"name\": \"/toolcall/baidu_search\"}}]}\n\ndata: {\"role\": \"tool\", \"status\": \"done\", \"content\": [{\"name\": \"/toolcall/crawl\", \"type\": \"text\", \"text\": {\"info\": \"### DeepSeek与主流大模型性能参数深度解析：技术选型指南\\n\\n**1.1 架构设计差异** \\nDeepSeek采用混合专家模型（MoE）架构...\"}, \"event\": {\"id\": \"16527614-1326-49b8-954f-09ab05f90a90\", \"status\": \"done\", \"name\": \"/toolcall/crawl\"}}]}\n\ndata: {\"role\": \"tool\", \"status\": \"done\", \"content\": [{\"name\": \"/toolcall/structured_outline\", \"type\": \"json\", \"text\": {\"data\": \"{\\\"thinking\\\":\\\"研究主题为‘AI市场调研报告’...\\\",\\\"title\\\":\\\"AI市场全景调研与发展趋势分析报告\\\",\\\"sub_chapters\\\":[{\\\"title\\\":\\\"AI市场宏观概览与规模分析\\\",\\\"sub_chapters\\\":[...]}]}\"}, \"event\": {\"id\": \"ded1a16c-9911-4d76-b731-84f1a23c8852\", \"status\": \"done\", \"name\": \"/toolcall/structured_outline\"}}]}
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
\n
","fields":{"slug":"nmwbl4w38","title":"深度研究Agent对话","date":"2026-11-11","extractedHeadings":[{"value":"权限说明","depth":2},{"value":"请求参数","depth":2},{"value":"示例代码","depth":2},{"value":"返回响应","depth":2}]},"headings":[]}},"pageContext":{"isCreatedByStatefulCreatePages":false,"slug":"nmwbl4w38","prev":{"id":"Nmwbl4w18","name":"通用对话响应参数","path":"Nmwbl4w18","filePath":"Agent/应用调用/官方Agent/通用对话响应参数.md","seo":null,"parentIds":["Mmwbl4qxa","Mmwbl4vtd","mmwbl4vtl"],"parents":[{"id":"Mmwbl4qxa","documentId":"14a689a9-e16b-4e40-ba67-9e0a3a66f51e","name":"Agent","repoName":"qianfan-api","filePath":"Agent","disabled":false,"path":"Mmwbl4qxa","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null},{"id":"Mmwbl4vtd","documentId":"02ac9724-255f-408f-90cd-907eaa1b9e82","name":"应用调用","repoName":"qianfan-api","filePath":"Agent/应用调用","disabled":false,"path":"Mmwbl4vtd","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null},{"id":"mmwbl4vtl","documentId":"27f07323-caf6-4ee0-b391-b2cdd9cfb059","name":"官方Agent","repoName":"qianfan-api","filePath":"Agent/应用调用/官方Agent","disabled":false,"path":"mmwbl4vtl","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null}]},"next":{"id":"Kmwbl4w5k","name":"自定义Agent","path":"Kmwbl4w5k","filePath":"Agent/应用调用/自定义Agent/新建对话.md","seo":null,"parentIds":["Mmwbl4qxa","Mmwbl4vtd","hmwbl4w5c"],"parents":[{"id":"Mmwbl4qxa","documentId":"14a689a9-e16b-4e40-ba67-9e0a3a66f51e","name":"Agent","repoName":"qianfan-api","filePath":"Agent","disabled":false,"path":"Mmwbl4qxa","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null},{"id":"Mmwbl4vtd","documentId":"02ac9724-255f-408f-90cd-907eaa1b9e82","name":"应用调用","repoName":"qianfan-api","filePath":"Agent/应用调用","disabled":false,"path":"Mmwbl4vtd","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null},{"id":"hmwbl4w5c","documentId":"35a7f0b8-fe64-4fa3-925f-01cbb6b30b0b","name":"自定义Agent","repoName":"qianfan-api","filePath":"Agent/应用调用/自定义Agent","disabled":false,"path":"hmwbl4w5c","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null}]},"parents":[{"id":"Mmwbl4qxa","documentId":"14a689a9-e16b-4e40-ba67-9e0a3a66f51e","name":"Agent","repoName":"qianfan-api","filePath":"Agent","disabled":false,"path":"Mmwbl4qxa","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null},{"id":"Mmwbl4vtd","documentId":"02ac9724-255f-408f-90cd-907eaa1b9e82","name":"应用调用","repoName":"qianfan-api","filePath":"Agent/应用调用","disabled":false,"path":"Mmwbl4vtd","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null},{"id":"mmwbl4vtl","documentId":"27f07323-caf6-4ee0-b391-b2cdd9cfb059","name":"官方Agent","repoName":"qianfan-api","filePath":"Agent/应用调用/官方Agent","disabled":false,"path":"mmwbl4vtl","lastMergeTime":null,"isApiDoc":null,"httpMethod":null,"seo":null,"sourceOrgName":null,"sourceRepoName":null,"sourceDocumentId":"","solutionTag":null,"solutionColor":null}],"specificSeo":null}}}