[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84394-en":3,"doc-seo-84394-105":29,"detail-sidebar-cat-0-en-105":90},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},84394,7971461741311,"Ophelia","https://ap-avatar.wpscdn.com/avatar/74000253aff267980c6?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779345379180704826",6,"Technology","WebSwarm Recursive Multi-Agent Orchestration for Deep-and-Wide Web Search","LLM-based web search agents shift information seeking from simple factoid QA to deep-and-wide, research-like investigation. However, single ReAct agents face limited context and a constrained long trajectory, while existing multi-agent approaches improve coverage but remain weak in recursive depth, collaboration adaptability, and evidence-grounded expansion. WebSwarm introduces a progressive recursive delegation framework that builds task decomposition, recursive expansion, and collaboration during inference by dynamically instantiating agentic nodes and propagating evidence upward. Experiments on multiple benchmarks show consistent gains and analyses explain the effectiveness.","WebSwarm: Recursive Multi-Agent Orchestration for Deep-and-Wide Web Search  \nXiaoshuai Song 1 ∗ , Liancheng Zhang 1 ∗ , Kangzhi Zhao2†, Yutao Zhu 1 , Zhongyuan Wang2∗, Guanting Dong 1 , Jinghan Yang2∗, Han Li2 , Kun Gai2 , Ji-Rong Wen 1 , Zhicheng Dou 1†  \n1 Gaoling School of Artificial Intelligence, Renmin University of China 2 Kuaishou Technology  \n{songxiaoshuai, [dou}@ruc.edu.cn](dou}@ruc.edu.cn), [kangzhi.zhao@outlook.com](kangzhi.zhao@outlook.com)  \narXiv :2607 .08662v 1 [ cs .CL] 9 Jul 2026  \nAbstract  \nLarge language model (LLM)-based web search agents are transforming information seeking from simple factoid question answering into complex, deep-and-wide search and research-oriented tasks. A single ReAct-style agent is constrained by one long trajectory and limited context, making it difficult to handle depth and coverage simultaneously. Existing multi-agent systems improve search coverage through parallel execution and aggregation, but still exhibit clear limitations in recursive depth, collaboration adaptability, and evidencegrounded expansion. We propose WebSwarm, a progressive recursive delegation framework that jointly constructs task decomposition, recursive expansion, and agent collaboration during inference. WebSwarm dynamically instantiates agentic search nodes, each coupling a local objective with a search mode that specifies how the node should organize search and collaboration. Each node can either solve its objective itself or further delegate child nodes; after solving, it returns evidence and results upward, enabling parent nodes to further expand, revise, or aggregate the search process. To guide this process, WebSwarm first probes how task-relevant information is organized on the web to ground subsequent node expansion, and reuses process-level experience across homogeneous sibling nodes. Experiments on BrowseComp-Plus, WideSearch, DeepWideSearch, and GISA show that WebSwarm consistently outperforms single-agent and multi-agent baselines on deep, wide, and interleaved deep-and-wide tasks. Further analyses of ablation, task difficulty, web tool efficiency, and model generalization explain WebSwarm’s effectiveness and provide insights for multi-agent search systems1 .  \nIntroduction  \nThe development of large language models (LLMs) is driving web information seeking toward agentic search, enabling search agents to autonomously perform multi-turn search and web browsing to gather information for user queries (Zhu et al. 2026a; Xi et al. 2025) . As this paradigm evolves, search agents are moving beyond simple factoid QA toward more complex information-seeking tasks, such as research-level search and report-oriented investigation. This requires agents to support both deep and wide search: deep search resolves multi-hop dependencies and constraints, while wide search maintains sufficient coverage across candidate entities, web  \n1Github: [https://github.com/songxiaoshuai/WebSwarm](https://github.com/songxiaoshuai/WebSwarm)  \n∗Work done during internship at Kuaishou, supervised by Kangzhi Zhao ([kangzhi.zhao@outlook.com](kangzhi.zhao@outlook.com)).  \n†Corresponding author.  \n\n| Vanilla\u003Cbr> |  |  |  | • Shallow root-level split\u003Cbr>• Fixed collaboration mode |\n| --- | --- | --- | --- | --- |\n|  | Agent 1\u003Cbr> |  |  |  |\n|  | Agent 2  |  |  |  |\n|  | Agent 3  |  |  | • Web-structure misaligned |\n\n\n| Plan-then-Orchestrate |  | • Plan before evidence\u003Cbr>• Static decomposition graph\u003Cbr>• Bottlenecked by initial plan |\n| --- | --- | --- |\n|  Plan\u003Cbr> |  T4  T1  T2  T6  A  T3  T5 \u003Cbr>\u003Cbr>\u003Cbr>\u003Cbr> |  |\n\nFigure 1: Illustration of representative multi-agent orchestration paradigms and WebSwarm.  \npages, and information sources. Recently, a series of benchmarks have evaluated the capability boundaries of search agents from the perspectives of depth, width, and their nested interaction (Wei et al. 2025; Chen et al. 2025b; Wong et al. 2026; Lan et al. 2025; Zhu et al. 2026b) .  \nBoth benchmark results and practical experien","cbCaieU1wSS23Ls8","https://ap.wps.com/l/cbCaieU1wSS23Ls8","pdf",1096779,1,19,"English","en",105,"# Abstract\n# Introduction\n## Deep-and-wide search challenges\n## Limitations of existing multi-agent systems\n# Proposed framework (WebSwarm)","[{\"question\":\"What problem does WebSwarm address in LLM-based web search agents?\",\"answer\":\"WebSwarm targets the difficulty of handling both depth and width simultaneously, where single-agent ReAct struggles and existing multi-agent systems lack sufficient recursive depth, flexible collaboration, and evidence-grounded expansion.\"},{\"question\":\"How does WebSwarm construct task decomposition and collaboration during inference?\",\"answer\":\"WebSwarm uses a progressive recursive delegation framework that dynamically instantiates agentic search nodes, each with a local objective and a search mode, then delegates to child nodes when needed.\"},{\"question\":\"How are evidence and results used within WebSwarm?\",\"answer\":\"After a node solves its objective or delegates further, it returns evidence and results upward, enabling parent nodes to expand, revise, or aggregate the search process.\"}]",1784195279,48,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":27},"webswarm-recursive-multi-agent-orchestration-for-deep-and-wide-web-search","",{"@graph":35,"@context":84},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/technology/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/webswarm-recursive-multi-agent-orchestration-for-deep-and-wide-web-search/84394/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What problem does WebSwarm address in LLM-based web search agents?","Question",{"text":74,"@type":75},"WebSwarm targets the difficulty of handling both depth and width simultaneously, where single-agent ReAct struggles and existing multi-agent systems lack sufficient recursive depth, flexible collaboration, and evidence-grounded expansion.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does WebSwarm construct task decomposition and collaboration during inference?",{"text":79,"@type":75},"WebSwarm uses a progressive recursive delegation framework that dynamically instantiates agentic search nodes, each with a local objective and a search mode, then delegates to child nodes when needed.",{"name":81,"@type":72,"acceptedAnswer":82},"How are evidence and results used within WebSwarm?",{"text":83,"@type":75},"After a node solves its objective or delegates further, it returns evidence and results upward, enabling parent nodes to expand, revise, or aggregate the search process.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,112,117,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":110,"slug":111},50,"technology",{"id":113,"doc_module":4,"doc_module_name":45,"category_name":114,"show_sort_weight":115,"slug":116},7,"Healthcare",40,"healthcare",{"id":118,"doc_module":4,"doc_module_name":45,"category_name":119,"show_sort_weight":120,"slug":121},8,"Research & Report",30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":105,"slug":136},"General","general"]