"Crawling Websites to Extract Data" matching MCP servers:

MCP server for Averra Extract — lets AI agents like Claude, Cursor, and ChatGPT convert any webpage into clean, LLM-ready Markdown.

Enables an AI assistant to reliably read and chunk PDF/text documents, validate extracted JSON against a schema with full error paths, and save structured output—all confined to a single allowed directory.

A paid MCP server that extracts structured product data (schema.org/Product) from any URL via headless browser, charging $0.01 USDC per call through x402 micropayments.

Provides deterministic, schema-driven structured data extraction from PDF documents via regex/heuristic matching, enabling repeatable, inspectable, and cost-free field extraction without LLM API calls.

Enables extracting precise data from web pages by parsing the DOM with CSS selectors and returning structured JSON, while also offering tools to inspect page structure, discover candidate selectors, and verify them without loading full pages into context.

Extracts any Instagram post, reel, or multi-slide carousel and converts captions, OCR text, and metadata into structured screener filters such as PER, PBV, ROE, RSI, and volume for reusable IG-to-screener workflows.

Enables structured document understanding of local PDFs via a dual-extract MCP server, combining MinerU text/layout and Qwen3-VL vision with fusion adjudication for field extraction, tables, formulas, and validation. Fully local and offline.

MCP server that turns video URLs into transcripts and important keyframes locally, with no cloud or API keys.

Enables OCR text extraction from images via URLs or base64, with pay-per-call using x402 micropayments (USDC on Base L2).

Enables small local models to persist and retrieve long-term user facts through remember and recall tools, using SQLite FTS5 keyword search for lightweight memory without external vector databases or heavy VRAM.

Extracts brand identity (colors, typography, spacing) from any website and generates HTML/CSS replicas or applies branding to templates.

FastMCP server for PDF text extraction and metadata retrieval, providing structured JSON output for pages, metadata, and specific page ranges.

An MCP server for safe web crawling and scraping of public and authenticated sites using Playwright and Crawlee. It provides tools for scraping single pages, crawling entire sites, validating sessions, and listing supported sites, with built-in security protections and adapter-based extensibility.

MCP server for the loc8n Geographic Data API. Exposes U.S. demographics, housing, mortgage, migration, employment, and geographic data as tools.

Enables local read-only access to TheRundown Product API data including sports, affiliates, market definitions, events, futures, and open main lines using your own API key.

AI-powered MCP server for file renaming, organization, and PDF splitting using natural language prompts and content analysis.

Enables searching and retrieving documentation from crawled documentation sites as an MCP server, allowing coding agents to query real docs instead of relying on training data.

Enables natural language search and discovery of open-access scientific datasets through the EOSC Data Commons OpenSearch service. Provides tools to search datasets and retrieve file metadata using LLM-assisted queries.

Enables agents to deterministically extract structured entities from input payloads and validate or coerce them against JSON schemas, returning a standardized response dossier with execution telemetry. It runs as a zero-dependency MCP server that plugs into Claude Desktop, Cursor, and other MCP clients.

Provides read-only access to open, verified vehicle brand, model, and specification data (Italy first, later EU) through MCP tools, enabling AI agents and apps to query the dataset programmatically.