The Haunt API server enables AI agents to perform structured web data extraction and monitor API usage via two tools:
web_extract: Provide a URL and a plain-English description of what data to extract, and receive clean JSON back. Automatically handles JavaScript rendering, Cloudflare protection, and complex page layouts — no browser management or parsing setup required.
get_usage: Check your current API consumption and remaining credits against your plan limits.
It integrates via MCP (Model Context Protocol) with Claude Desktop, Cursor, VS Code, or Windsurf, using a simple API key passed as a parameter on each call.
Handles Cloudflare-protected websites automatically during web extraction, bypassing anti-bot measures and JavaScript rendering challenges to extract structured data.
Provides REST API access for web extraction with the same functionality as the MCP tools, allowing HTTP clients to extract structured data from URLs using prompts and API keys.
Eliminates the need for Docker containers or local infrastructure by providing fully hosted web extraction capabilities through the MCP server.
Automatically handles JavaScript rendering during web extraction, processing complex dynamic websites to extract structured data from pages that require JavaScript execution.
This repository has moved to Darko893/mcp-server.
That is the canonical source for the @hauntapi/mcp-server package and the
hosted Haunt MCP server.
Hosted MCP endpoint: https://hauntapi.com/mcp/server
This repository is archived and is no longer updated.
Extract the main article content from a news article or blog post. Returns title, body text, author, and publish date as structured JSON. Handles paywalled and JavaScript-rendered articles. Optimised for editorial content — use extract_url for product pages, listings, or generic data. Read-only — makes no changes to any external system. Requires HAUNT_API_KEY environment variable. Free tier: 100 requests/month. Returns an error if rate limit or API key is invalid.
Inputs: url (required).
Pull metadata from any URL: title, description, Open Graph tags, Twitter cards, canonical URL. Returns structured JSON with all available meta information. Useful for link previews, SEO analysis, and content categorisation. Use extract_url for page body content or extract_article for full articles. Read-only — makes no changes to any external system. Requires HAUNT_API_KEY environment variable. Free tier: 100 requests/month. Returns an error if rate limit or API key is invalid.
Inputs: url (required).
Extract structured data from any web page by providing a URL and describing what you want. Returns clean JSON with exactly the fields you asked for — no HTML parsing needed. Handles JavaScript-rendered pages and Cloudflare-protected sites automatically. This is the general-purpose extraction tool. Use extract_article for full article content or extract_metadata for page meta tags — they are optimised shortcuts. Read-only — makes no changes to any external system. Requires HAUNT_API_KEY environment variable. Free tier: 100 requests/month. Returns an error if rate limit or API key is invalid.
Inputs: url (required), prompt (required).
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 3 tools
The tools have overlapping purposes, as all three extract data from URLs, with extract_url described as a general-purpose tool that can handle what the other two do. However, the descriptions clearly differentiate their optimised use cases: extract_article for editorial content, extract_metadata for meta tags, and extract_url for custom extractions, which helps reduce confusion but still leaves some ambiguity.
All tool names follow a consistent verb_noun pattern with 'extract_' as the prefix, using snake_case uniformly. This predictability makes it easy for agents to understand the naming convention and relate the tools to their shared purpose of extraction.
With 3 tools, the count is reasonable for a focused extraction server, but it feels slightly thin as it only covers read-only extraction operations without broader management or configuration tools. The tools are well-scoped within their domain, but the set could benefit from additional complementary functions.
The server covers core extraction tasks for URLs, articles, and metadata, but there are notable gaps such as no update, delete, or management operations (e.g., for API keys or usage tracking). As a read-only extraction service, it lacks lifecycle coverage, which limits agent workflows to just data retrieval without any control or monitoring capabilities.
Web scraping, extraction and search for AI agents. Hosted MCP with workspace API keys.
Web scraping, browser automation, screenshots and AI data extraction for AI agents via MCP.
Web scraping for agents. Point it at a URL and it returns the page as clean markdown, JavaScript-rendered pages included. Point it at a site and it maps the URLs or crawls the section you need in the background, a few pages at a time so results fit in the conversation. Search the web and read full pages, extract fields with a JSON schema you define (validated, never invented), read a store's catalogue or a blog's posts from the platform's own feed, and check whether a page has changed. Failed requests cost nothing. The free plan includes 1,500 credits a month.
Web scraping for AI agents. Extract text and metadata from any URL worldwide. $0.005/page.
