qwen-web-research-mcp
This server is an MCP tool for web research: it fetches web pages and uses a local Ollama LLM to answer questions about their content.
analyze_page(url, question) — Fetches a given URL, extracts the full page content, and answers a question about it; long pages are auto-split and analyzed with map-reduce so nothing is skipped.
search_site_and_analyze(site, phrase, question, max_results) — Searches a site via DuckDuckGo's
site:operator for pages containing a phrase, then analyzes each matching page to answer the question.list_available_models() — Lists the Ollama models available on the configured server.
Works with any Ollama model (default
qwen3:14b), not tied to a specific model family.Robust content extraction via trafilatura, independent of site layout/CSS selectors.
Supports multiple MCP transports: stdio, SSE, and streamable-http.
Configurable via environment variables for Ollama host/port/model and HTTP bind settings.
Can run locally or via Docker; integrates easily with MCP clients like Open WebUI.
Reports progress during long operations to prevent client timeouts.
Integrates with DuckDuckGo's site search to find pages on a given domain containing a phrase.
Uses local Qwen models via Ollama to extract answers and analyze page content.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@qwen-web-research-mcp analyze https://en.wikipedia.org/wiki/Artificial_intelligence and summarize key points"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
web-research-mcp
MCP server for web research: given a URL (or a site + phrase to search), it fetches the full page content, and uses a local LLM (via Ollama — any model you have pulled, not tied to a specific model family) to extract/answer a question about it. Long pages are automatically split into chunks and analyzed with a map-reduce strategy, so nothing is skipped.
Content extraction uses trafilatura, which strips navigation/ads/boilerplate based on content heuristics rather than site-specific CSS selectors — so it keeps working if a site's layout changes. Site-scoped search uses DuckDuckGo's site: operator, so any domain works without custom scraping code per site.
Tools
analyze_page(url, question)— fetch a page and answerquestionabout its full content.search_site_and_analyze(site, phrase, question, max_results)— find pages onsitecontainingphrase, then runanalyze_page-style extraction on each match.list_available_models()— list the models available on the configured Ollama server.
Related MCP server: searxng-mcp
Requirements
Python 3.11+
An Ollama server with at least one model pulled (defaults to
qwen3:14b, but any Ollama model works)
Configuration (environment variables)
Variable | Default | Description |
|
| Ollama model to use |
|
| Port of the Ollama server |
| auto-detected | Override the Ollama host IP. If unset, it's read from |
|
|
|
|
| Bind host (HTTP transports only) |
|
| Bind port (HTTP transports only) |
Run locally
python3 -m venv .venv
.venv/bin/pip install -e .
.venv/bin/python -m web_research.serverRun with Docker
docker build -t web-research-mcp .
docker run -p 8001:8000 \
-e MCP_TRANSPORT=streamable-http \
-e MCP_HOST=0.0.0.0 \
-e OLLAMA_HOST_IP=host.docker.internal \
web-research-mcpOr as a service in a docker-compose.yml alongside other tools (e.g. Open WebUI):
web-research:
build: ./web-research-mcp
ports:
- "8001:8000"
extra_hosts:
- "host.docker.internal:host-gateway"
environment:
- MCP_TRANSPORT=streamable-http
- MCP_HOST=0.0.0.0
- MCP_PORT=8000
- OLLAMA_HOST_IP=host.docker.internal
- OLLAMA_MODEL=qwen3:14bThen point your MCP client (e.g. Open WebUI's Tools/Connections settings) at http://web-research:8000/mcp (internal Docker network) or http://localhost:8001/mcp (from the host).
Available Tools
3 toolsanalyze_pageA
Fetch a web page, and use a local Qwen model to extract/answer question
about its full content. Long pages are automatically split into chunks and
analyzed piece by piece (map-reduce), so no content is skipped.
Reports progress while working (page fetch retries, per-chunk analysis), so a client that respects MCP progress notifications won't time out waiting on a long page -- this can take 30s-3min or more.
Args: url: The page to fetch and analyze. question: What to extract or answer about the page's content.
Inputs: url (required), question (required). Returns: result
list_available_modelsA
List the Qwen/Ollama models currently available on the local Ollama server.
Inputs: no named parameters. Returns: result
search_site_and_analyzeA
Search a specific site for pages/listings containing phrase, then run
analyze_page-style extraction on each match to answer question.
Works on any site without site-specific scraping code: it uses DuckDuckGo's
site: search to find matches, then trafilatura + Qwen to read and filter
each page's content, so it keeps working even if the site's layout changes.
Each match takes roughly 30s-3min to fetch and analyze (more for long pages), so a large max_results will take proportionally long to return. Progress is reported per match and per chunk, for MCP clients that respect it.
Args: site: Domain to search within, e.g. "example.com". phrase: Exact phrase the publication/listing must contain. question: What information to extract from each matching page. max_results: Max number of matching pages to analyze (default 5).
Inputs: site (required), phrase (required), question (required), max_results. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 3 tools
The two analysis tools are cleanly separated by input: analyze_page targets one explicit URL, while search_site_and_analyze first discovers matching pages on a domain and then analyzes them. list_available_models is clearly a separate utility, so there is no real ambiguity between tools.
All names use snake_case and lead with a verb, making the set predictable and readable. The main deviation is that search_site_and_analyze combines two actions into one name, while the other tools use a simpler verb_noun pattern.
Three tools is small but defensible for a focused web-research helper: two research workflows plus a model discovery utility. It is slightly minimal, but there is no obvious bloat or unnecessary duplication.
The server covers single-page analysis and site-scoped search, which are useful core pieces. However, it lacks a general web-search flow or multi-site research capability, so open-ended web research requests would have no obvious tool to use.
Maintenance
Related MCP Connectors
One MCP server for all web data. Scrape, search, crawl, and AI research in one install.
Turn any public website into an MCP server for agents to search, read and navigate.
Hosted Web Crawler MCP server for Claude, ChatGPT and Cursor, with each action switched per account.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceMCP server for local semantic search over web content, enabling AI agents to ingest, index, and query pages with hybrid retrieval and token budget control.4GPL 2.0
- AlicenseAqualityBmaintenanceMCP server for private web search via self-hosted SearXNG with local reranking, full-page content fetching via Firecrawl, and optional Ollama-powered query expansion and summaries.7609 npm29MIT
- AlicenseNot gradedqualityCmaintenanceA self-contained web-research MCP server that lets local LLM agents search, fetch, and synthesize web content using tools like web_search, web_fetch, and web_research.2MIT
- AlicenseNot gradedqualityDmaintenanceSemantic web search MCP server that navigates the live web using a cheap LLM and local embeddings, caching results for fast responses to similar queries without relying on general search engines.MIT