Thunderbit, an AI web data platform based in San Francisco, has launched its developer API, Model Context Protocol (MCP) server, and Command Line Interface (CLI) to enhance web data extraction capabilities. These tools allow developers to transform complex websites into clean Markdown or structured data, which is crucial for AI agents and automation workflows. Central to this launch is Thunderbit Distill, an adaptive HTML-to-Markdown engine that excels in converting various web page types without needing site-specific rules. In tests, Distill achieved a ROUGE-L score of 0.87, indicating its effectiveness in producing high-quality Markdown. Additionally, the new Extract feature enables users to return structured JSON or CSV data from URLs based on user-defined schemas. This development aims to provide more reliable and cleaner data for AI applications, reducing the reliance on traditional scraping methods that can fail with layout changes. Thunderbit's tools are now available for developers, extending the platform's capabilities beyond its no-code offerings, which are already popular among sales and research teams.