1386 Model Context Protocol servers — searchable, rated, ready to install.
Web scraping MCP server — scrape, extract structured data, screenshot any site with anti-bot bypass.
Create and control browser sessions for web automation, testing, and AI agent tasks
Scrape Google search results with SERP data, ads, and knowledge panels
Local Playwright browser through residential proxies — agent picks country + sticky identity
Crawl, scrape, search the web, and automate browsers at scale with anti-bot bypass.
macOS desktop control for Claude: click, type, screenshot, AX tree, browser control, agentic loops.
Undetectable browser automation for LLM agents, spoken over MCP.
Track brand mentions & keywords on Bluesky. Sentiment, engagement, author reach. Pay per result.
Figma MCP server for AI agents. 9 tools: design context, variables, screenshots, and FigJam.
Generate PDFs, screenshots, and documents via any MCP-compatible AI agent.
Scrape the FAA Aircraft Registry by N-number — aircraft make, model, year, owner name, address, airw
Reddit browser for AI assistants. Browse without API keys; add credentials for search and analysis.
Lightweight browser automation for AI agents: navigate, screenshot, extract, PDF. One small binary.
AI-native browser for agents. 71 tools with self-healing, semantic extraction, vault CLI.
Outscraper MCP business discovery, Maps intelligence, enrichment, reviews, and contact data.
Playwright MCP-compatible browser automation bridge for CloakBrowser Chromium.
Give AI agents clean, LLM-ready web data — scrape any URL to markdown or extract structured JSON.
Scrape Amazon products, search, and async batch ASIN lookups across 20 marketplaces
use any website on a user's behalf in a cloud browser: navigate, sign in, fill forms, buy, download.
980+ AI tools as MCP servers. Finance, lead gen, scraping, dev tools, media, research, and more.
Scrape any website, extract structured data, and collect web content at scale with AI agents
MCP server that drives the user's real macOS Chrome via AppleScript. 32 tools.
Headless browser primitives for AI agents when sites need real JS rendering.
Generic URL crawl + HTML extraction — fallback for sites without dedicated MCPs.
POST /mcp