One MCP server. All web data. Search, scrape, crawl, proxy, and AI research — in a single npx command.
| Tool name | Description | Destructive? |
|---|---|---|
| novada_search | Search the web via Google, Bing, DuckDuckGo, Yahoo, or Yandex. Returns structured results with titles, URLs, snippets. S | ✓ no |
| novada_extract | Extract main content, title, description, and same-domain links from one or multiple URLs (batch mode). Supports markdow | ✓ no |
| novada_crawl | Crawl a website using BFS or DFS, extract content from multiple pages concurrently. Supports select_paths and exclude_pa | ✓ no |
| novada_map | Discover all URLs on a website without extracting content. BFS link collection with path-diverse queuing, max_depth cont | ✓ no |
| novada_research | Multi-step research: generates 3-10 parallel search queries, deduplicates up to 15 unique sources, returns cited report. | ✓ no |
| novada_proxy | Get Novada residential proxy configuration for use in HTTP clients, curl, or shell scripts. Returns proxy URL, shell exp | ✓ no |
| novada_scrape | Use when you need structured data from a specific platform — not raw HTML, but clean tabular records. Supports 129 platf | ✓ no |
| novada_verify | Use when you have a factual claim and need to check if it's supported by web sources. Runs 3 parallel searches (supporti | ✓ no |
| novada_unblock | Use when you need the raw rendered HTML of a blocked or JS-heavy page. Forces JS rendering via Web Unblocker or Browser | ✓ no |
| novada_browser | Use when you need to interact with a web page — click buttons, fill forms, scroll, take screenshots, or execute JavaScri | ✓ no |
| novada_health | Check which Novada API products are active on your API key. Best for: First-time setup, diagnosing why a tool is failing | ✓ no |