SuperAPI AI Directory
Partner with us中文AI newsFavorites
All toolsAI developer platformsidapixl-web-research-mcp

idapixl-web-research-mcp

Web research on Apify: Brave and DuckDuckGo search, page fetching to clean Markdown, and multi-step research with relevance scoring and key fact extraction.

Open website

About this tool

Purpose

idapixl-web-research-mcp is an MCP (Model Context Protocol) server designed for AI agents that need to search the web, read web pages, and synthesize research results. It integrates search, scraping, and synthesized research into a single tool call, enabling AI agents to complete the full process from retrieval to information distillation.

Main Capabilities

The server provides three tools that work together:

  1. web_search: Performs a web search. Parameters include query (required, search query, up to 500 characters), maxResults (optional, 1–10, default 5), freshness (optional, time filter: pd past day, pw past week, pm past month, py past year, supported only by Brave).
  2. fetch_page: Fetches and reads a page, converting it into clean Markdown. Parameters include url (required, page URL), timeoutSecs (optional, 5–120, default 30), extractMode (optional, article extracts the main body content (default), full extracts the entire body, links extracts all links).
  3. research_topic: Performs multi-step deep research. Parameters include topic (required, research topic or question, up to 500 characters), maxSources (optional, 1–5, default 3), focus (optional, a sub-question used to narrow the research scope, such as "performance benchmarks", up to 200 characters). This tool scores sources for relevance, deduplicates near-duplicate sources (filtering content with more than 70% sentence overlap), and extracts key sentences containing query terms.

How to Use

Add the server to your MCP client configuration (such as Claude Desktop, Cursor, etc.), and replace YOUR_APIFY_API_TOKEN with a token from apify.com/account/integrations. For Cursor / Windsurf, add it via Settings > MCP Servers > HTTP URL.

Brave Search API setup (optional but recommended): By default, DuckDuckGo HTML scraping is used, and it works without any key. For better results and time filtering, you can sign up for the free tier at brave.com/search/api (2,000 queries per month), copy the API key, and set the Brave Search API Key in the actor input (stored as a secret and not exposed in logs). After setting the key, web_search and research_topic will automatically use Brave Search and fall back to DuckDuckGo on error.

Use Cases

Suitable for scenarios where AI agents need online information retrieval, web page content reading, and multi-source synthesized research, such as data collection, fact-checking, and competitor research. Supports MCP clients such as Claude Desktop, Cursor, and Windsurf.

Limitations and Billing

  • Pay-per-use, charged only for tool calls, with no subscription and no idle fees.
  • Charges are deducted before execution; if a page fails to load after the charge, the tool returns an error.
  • Stateless: each MCP request gets an isolated server instance.
  • Using Apify Residential proxies can improve the success rate for pages that block data center IPs.
  • Tech stack: TypeScript, MCP SDK 1.x, Apify SDK 3.x.
  • Built by Idapixl.

Related tools

Browse category