Search lenses
Lets you restrict searches to curated source lists, so the AI searches the sites you trust instead of the open web by default.
web-researcher-mcp
web-researcher-mcp is an open-source MCP server that adds web search, source reading, and cited research workflows to Claude, Cursor, and other MCP clients.
Overview
web-researcher-mcp is an open-source Model Context Protocol server for AI-assisted research. It connects tools like Claude, Cursor, and other MCP clients to web search and source reading so the assistant can gather evidence instead of relying on unsourced answers.
The repository positions the tool around source quality and traceability. It lets users narrow searches to curated “search lenses” for specific domains, read full documents, and attach citations so research can be reproduced and checked later.
Core features
Broad source coverage
Supports research across papers, patents, SEC filings, US court records, economic data, news, web pages, images, PDFs, Word docs, YouTube transcripts, and Hacker News threads.
Full-text reading
Reads full source content rather than only snippets, which helps the model ground answers in complete documents.
Formatted citations
Returns citations in standard formats such as APA and MLA, with links that are intended to be directly usable.
Local execution and privacy
Runs locally on the user’s machine, with the README stating that queries stay private and are not seen by the vendor.
Fallback search engines
Provides multiple search engines as backup so that if one search provider has problems, others can continue the workflow.
Common use cases
Academic research
Use curated source sets to keep a literature review focused on trusted journals, databases, or institutional sites instead of the wider web.
Business research
Pull source-backed evidence for deliverables that need checkable citations, such as client reports, internal memos, and competitive analysis.
Legal research
Search court records, filings, and related documents while keeping a reproducible record of where each result came from.
Medical research
Look up medical papers, clinical references, and other high-trust sources when accuracy matters more than broad summarization.
Private research workflows
Use the local, private workflow to research sensitive topics without sending queries through a third-party hosted research app.
Pros and Cons
Pros
- Works with existing MCP clients rather than requiring a separate research app.
- Supports curated source selection through search lenses, which helps narrow research to trusted domains.
- Reads full documents instead of relying only on result snippets.
- Provides cited outputs in standard formats and keeps a paper trail of searches.
- Runs locally, which the repository says keeps queries private.
Cons
- Setup is spread across several install paths, including `uvx`, Homebrew, shell scripts, PowerShell, and `.mcpb` bundles.
- Some richer search features need a key or extra configuration beyond the default DuckDuckGo setup.
FAQ
How do you install or connect web-researcher-mcp?
It can run through the `uvx web-researcher-mcp` command, a Homebrew install, a direct shell script installer, a Windows PowerShell installer, or a Claude Desktop `.mcpb` bundle. The README also says it can be connected to Claude Code automatically when the `claude` CLI is present.
Which AI clients does it work with?
The README says it works with Claude, Claude Desktop, Cursor, LM Studio, and any MCP client that supports tool use. It also notes that custom per-app configuration is available for other clients.
What kind of research workflows does it support?
The product is built to search the web, read full articles and PDFs, find papers and patents, and return grounded answers with citations. It also supports search lenses, which let you narrow searches to trusted source sets such as PubMed, SEC.gov, or arXiv.
Does it require an API key?
The README says DuckDuckGo web search works with no API key for a quick start. Richer providers and features such as `image_search` and `news_search` need a key, and configuration determines the search provider.
How does it handle privacy and traceability?
The repository states that it runs on your machine and keeps queries private. It also logs searches so you can reproduce the research process later.
Quick Facts
- Category
- Developer Tool
- Product type
- Open-source MCP server
- Primary users
- Researchers, analysts, lawyers, journalists, students, and enterprise teams
- Platform
- Works with Claude, Cursor, Claude Desktop, LM Studio, and other MCP clients
- Source domain
- github.com
- Pricing
- Free forever (open source)
web-researcher-mcp Alternatives
Hype
Hype is a web tool for finding trending YouTube topics by category, time range, and scoring mode. Spot emerging ideas and review source videos.
Ably Chat
Ably Chat is a chat API platform for building custom realtime chat applications. It supports room-based messaging, typing indicators, presence, reactions, and message updates, with usage-based pricing options for different deployment stages.
ClayHog
ClayHog is an AI search visibility platform for tracking brand presence in ChatGPT, Gemini, Perplexity, Claude, and Google AI Overviews from one dashboard.
Grok AI Assistant
Grok is a free AI assistant developed by xAI, engineered to prioritize truth and objectivity while offering advanced capabilities like real-time information access and image generation.
Ghost
Ghost is a terminal-based AI assistant for chatting, code generation, and CLI tasks. Includes free models, supports Linux, macOS, Windows, and is open source.
Paperpal
Paperpal 是一款面向学术写作的 AI 工具，提供文献阅读、语言润色、引用、查重、AI 检测和投稿前检查。它适合需要在英文论文和学术材料上进行写作、修改与合规自检的学生、研究人员和教师。