# Do you deliver data in an LLM-optimized format?

Yes, the Model Context Protocol (MCP) support is a built-in feature of our [Web Scraper API](https://oxylabs.io/products/scraper-api/web). The MCP process transforms raw HTML into formats that Claude, GPT, and other models can immediately understand. No manual reformatting is required.

{% hint style="info" %}
🛠️ MCP acts as a wrapper that structures the Web Scraper API’s output with context, metadata, and instructions, ensuring compatibility with LLMs.
{% endhint %}

Web Scraper API can also return [Markdown output](https://developers.oxylabs.io/scraping-solutions/web-scraper-api/features/result-processing-and-storage/output-types/markdown-output) by adding `"markdown": true` to your request. Markdown is a lightweight, LLM-friendly, and easy-to-read format that preserves structure and emphasis, making it ideal for chunking, metadata extraction, and preprocessing in AI workflows.

Alternatively, you can use [AI Studio](https://aistudio.oxylabs.io/) to scrape websites using natural language prompts. It can deliver content in Markdown format and parse data according to your prompt or a provided Pydantic or JSON schema.