Search

Search queries the web and returns ranked results with titles and snippets. Add scrapeOptions and every result is also scraped, so you get full markdown for each hit in one round trip. 5 credits per 10 results (rounded up, and at least 5 per search, since a search that finds nothing still queries the index), plus normal scrape credits for each result scraped successfully. Results are scraped within a 45-second budget; any that fail to scrape, or aren't reached in time, come back without a document and cost nothing extra. Identical queries are cached for about 10 minutes.

country is a two-letter code such as us or de, and language a language code such as en or pt-BR. A query the search provider refuses comes back as invalid_request at no charge.

Example#

curl -X POST "https://api.vermin.dev/v1/search" \
  -H "Authorization: Bearer $VERMIN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "open source web scrapers",
    "limit": 5,
    "scrapeOptions": { "formats": ["markdown"] }
  }'

Request#

Response#

{
  "success": true,
  "results": [
    {
      "position": 1,
      "url": "https://github.com/topics/web-scraping",
      "title": "web-scraping · GitHub Topics",
      "snippet": "Web scraping is the process of extracting data from websites…",
      "source": "GitHub",
      "document": { "url": "https://github.com/topics/web-scraping", "markdown": "# web-scraping\n…", "tier": "fetch" }
    }
  ],
  "credits_used": 10,
  "request_id": "req_01J9V6C8LM"
}

Here 5 results cost 5 credits, plus 5 successful scrapes at 1 each.