Copy, paste, haunt

Extraction recipes that prove the product.

These recipes focus on Haunt's best commercial lanes: company intelligence, docs, GitHub repositories, products, pricing, Reddit, and Hacker News. No fake social scraping cosplay.

Company pages

Company website to JSON

Use this for lead research, competitor positioning, vendor pages, and public business metadata.

curl https://hauntapi.com/v1/extract \
  -H "Content-Type: application/json" \
  -H "X-API-Key: PASTE" \
  -d '{
    "url": "https://example.com",
    "prompt": "Extract the company name, positioning, product categories, pricing link, and contact page link. Use null for missing fields."
  }'

GitHub

GitHub repository intelligence

Prefer the dedicated route for repo metadata. It is cheaper and cleaner than generic page extraction.

curl "https://hauntapi.com/v1/github/repo?owner=browserbase&repo=stagehand" \
  -H "X-API-Key: PASTE"

Products

Product or pricing page extraction

Good for public plan tables, pricing cards, product lists, and competitor packaging checks.

curl https://hauntapi.com/v1/extract \
  -H "Content-Type: application/json" \
  -H "X-API-Key: PASTE" \
  -d '{
    "url": "https://example.com/pricing",
    "prompt": "Extract plan names, prices, monthly limits, and highlighted features as structured JSON."
  }'

Docs

Docs page extraction

Turn docs and API references into concise structured facts for agents, QA, and onboarding.

curl https://hauntapi.com/v1/extract \
  -H "Content-Type: application/json" \
  -H "X-API-Key: PASTE" \
  -d '{
    "url": "https://docs.example.com/quickstart",
    "prompt": "Extract the install command, authentication steps, required environment variables, and first API call."
  }'

Reddit thread extraction

Use Haunt's dedicated Reddit route. Generic Reddit page scraping is intentionally not the reliable path.

curl https://hauntapi.com/v1/reddit \
  -H "Content-Type: application/json" \
  -H "X-API-Key: PASTE" \
  -d '{"subreddit":"webscraping","sort":"hot","limit":5}'

Hacker News

Hacker News item extraction

Use the public item route for story metadata, points, comment count, author, and source URL.

curl https://hauntapi.com/v1/hackernews/item/40310896 \
  -H "X-API-Key: PASTE"

Use What Haunt can extract before promising a customer anything weird. If your visitor is an agent builder, send them to the focused AI agent web extraction path so they can try the demo, get a key, and make one first call.

Ready to try it?

Start with the fixed demo response, then get a free key and make one safe first extraction.

Try demo response Get free key First-call guide