CoSearchAgent: A Lightweight Collaborative Search Agent with Large Language Models
| Field | Detail | |-------|--------| | Title | CoSearchAgent: A Lightweight Collaborative Search Agent with Large Language Models | | Authors | Peiyuan Gong, Jiamian Li, Jiaxin Mao | | Published | 2024-02-09 | | Source | https://arxiv.org/abs/2402.06360 | | Type | Demo paper (arXiv) |
TL;DR
CoSearchAgent is an LLM-powered agent that supports collaborative search during multi-party conversations. It ships as a Slack plugin and answers user queries with answers grounded in live web search results, asking clarifying questions when the information need is ambiguous.
Background and Motivation
Collaborative search supports multiple users working together to accomplish a specific search task. Research has found that designing lightweight collaborative search plugins within instant messaging platforms aligns better with users' collaborative habits. However, due to the complexity of multi-user interaction scenarios, it is challenging to implement a fully functioning lightweight collaborative search system. Therefore, previous studies on lightweight collaborative search had to rely on the Wizard of Oz paradigm.
In recent years, large language models (LLMs) have been demonstrated to interact naturally with users and achieve complex information-seeking tasks through LLM-based agents. This opens the door to building a real, working collaborative search system without human simulation.
What CoSearchAgent Does
To better support research in collaborative search, the authors propose CoSearchAgent, a lightweight collaborative search agent powered by LLMs. Key capabilities:
- Slack integration: designed as a Slack plugin, supporting collaborative search during multi-party conversations on the platform.
- Multi-user context understanding: understands queries and context within multi-user conversations.
- Grounded answers: searches the Web for relevant information via APIs and responds with answers grounded on the relevant search results.
- Clarification behavior: asks clarifying questions when information needs are unclear.
- A Survey of Conversational Search (ACM, Sep 2025)
- Engineering Conversational Search Systems (arXiv:2407.00997)
- ChatRetriever: Adapting LLMs for Generalized and Robust Conversational Dense Retrieval (arXiv:2404.13556)
- ConvGQR: Generative Query Reformulation for Conversational Search (arXiv:2305.15645)
- Original paper: CoSearchAgent: A Lightweight Collaborative Search Agent with Large Language Models — https://arxiv.org/abs/2402.06360
Significance
The proposed CoSearchAgent is highly flexible and useful for supporting further research on collaborative search. It replaces the Wizard of Oz setup with a deployable LLM-based agent, lowering the barrier for studying real multi-user search interactions. The code and demo video are accessible via the paper's arXiv page.