Web scraping API for AI agents

Turn a URL into useful context for an agent. Request the outputs you need in one call, rather than parsing an entire page in your application.

Extract Markdown from a URL

curl -i https://api.oassis.dev/web/v1/scrape \
  -H 'Content-Type: application/json' \
  -d '{"url":"https://example.com","formats":["markdown","links"]}'

The HTTP request requires payment. Its response groups the requested outputs; consult OpenAPI for the complete response schema.

Choose the output for the next step

• markdown: readable text for summaries and retrieval.
• html and links: markup and link discovery.
• controls and accessibility: page structure for agent interaction.
• screenshot and pdf: visual outputs.
• json: structured extraction with a prompt or schema.

Extract structured fields

curl -i https://api.oassis.dev/web/v1/scrape \
  -H 'Content-Type: application/json' \
  -d '{"url":"https://example.com","formats":["json"],"json":{"prompt":"Extract the page title and a short description."}}'

Prices and limits

On the HTTP API, a fresh Markdown output costs $0.002, JSON extraction costs $0.005, and a cached output costs $0.0002. Multiple outputs add together. Supported PDF, Word, Excel and CSV URLs use document conversion, priced at $0.002. Read the payment challenge for the exact request price.

A page must be publicly reachable. Access restrictions, page changes and unsupported content can prevent extraction; a browser tool does not guarantee access to every website.

Read a list or a whole site

Use web_scrape_batch for a known list of URLs. Use the crawling API when links must be discovered as you go.

Continue with MCP integration, x402 payments or the OpenAPI reference.