–

Search API that makes it easy and fast to scrape Google and other search engines.

This listing is not live yet. It will be published on August 8, 2026.

Haunt Web Extractor

Development

Visit website

Extract structured JSON from web pages using natural language prompts.

Haunt Web Extractor's website screenshot

About Haunt Web Extractor

Haunt Web Extractor is a web extraction API that converts public web pages into structured JSON or clean Markdown. It handles fetching, JavaScript rendering, and extraction through a single endpoint, and returns schema-checked structured data along with a confidence trace. When a page is blocked, empty, or login-only, the API returns a clear error (such as captcha_required) rather than returning fabricated data; it does not solve CAPTCHAs or bypass login walls.

The product is built for developers and AI agents that need reliable facts from web pages instead of raw HTML. It supports use cases including pricing monitoring (extracting plans, prices, limits and features from changing pages), company summarization (metadata and competitor context from public pages), and feeding structured or Markdown page data to AI agents and RAG pipelines. It is accessible via REST and MCP, with an SDK also mentioned; REST supports a response_format of "markdown", and the MCP package exposes an extract_markdown function for agent workflows. Page content is not stored, and extracted results are retained only if the user opts in.

Pricing is based on monthly credit plans: Free (£0/mo, 1,000 credits/month), Starter (£19/mo, 10,000 credits/month, adds residential proxy retries for blocked pages and scripted browser steps), Pro (£49/mo, 30,000 credits/month, adds custom headers and cookies in beta and a higher rate limit), and Scale (£99/mo, 80,000 credits/month, adds the highest rate limit and concurrency). Failed extractions do not consume credits, and there are no automatic overages. Credit costs vary by extraction type: simple pages use 1 credit, normal structured extraction 2 credits, browser-rendered pages 4 credits, and heavy pages 8 credits, with an additional credit per roughly 10,000 tokens of page content on very large pages, up to 3 extra credits.

Key features

Extracts structured JSON or clean Markdown from public web pages
Renders JavaScript-heavy pages before extraction
Returns clear errors instead of fake success on blocked, empty, or login-only pages
Residential proxy retries for blocked pages and scripted browser steps on paid plans
REST API and MCP support, including a markdown response format and extract_markdown for agents
Failed extractions do not consume credits

Frequently asked questions

Is there a free tier?

Yes, the Free plan gives 1,000 credits per month with no credit card required.

How much does it cost?

Plans are £0/mo for Free, £19/mo for Starter (10,000 credits), £49/mo for Pro (30,000 credits), and £99/mo for Scale (80,000 credits).

Can I use it with AI agents or MCP?

Yes, Haunt supports REST with a markdown response format and provides an MCP package with an extract_markdown function for agent workflows.

What happens if a page is blocked or behind a login?

Haunt detects many CAPTCHA and verification blocks and returns a clear error like captcha_required, but it does not solve CAPTCHAs or bypass login walls.

How do credits get used?

Simple pages typically use 1 credit, normal structured extraction 2, browser-rendered pages 4, and heavy pages 8, with failed extractions costing nothing.

Advertise here