1 request is counted as one successfully processed webpage or PDF. We don't charge you additionally for things like GB of data transfer or proxies—all those costs are included in the simple cost per request.

No, we only charge for successful requests. If you are using an endpoint that needs to make LLM calls (like `llm_extract`), we will pass down those third-party costs, but on our end, a request credit is only used upon success.

Because it's reliable (99.5% success rate), cost-effective (up to 70% cheaper), scalable, and flexible. Olostep is the only platform where you can create your own parsers to return deterministic, structured results at scale. Plus, you can get your free API keys here to test it yourself.

Absolutely. Our API infrastructure is built to scale and can handle billions of requests per month.

The API returns a request ID, and the data in your chosen format (Markdown, HTML, or structured JSON via Parsers or LLM extraction). Our `/answers` endpoint provides a direct answer, a JSON object in your defined schema, and the sources used for verification.

Yes. Use the `/scrapes` endpoint with the `llm_extract` parameter to describe what you want from a specific URL. For more complex tasks like multi-page navigation and validation, use the `/agents` endpoint to automatically find and extract data based on your prompt.