CADCrawl is a powerful AI-driven tool designed to help engineers and CAD professionals efficiently search and retrieve old CAD files such as STEP, PDF, and DXF. This tool is particularly beneficial for those working with extensive CAD archives, enabling them to find reusable work quickly and avoid unnecessary redraws. Key features include a plain-English search interface that allows users to search by job context, geometry, or customer details rather than exact filenames. CADCrawl uses a hybrid retrieval and reranking system that considers semantic meaning, filename relevance, folder context, extracted text, and file-type usefulness to surface the most relevant files. Built for Windows-heavy CAD workflows, it indexes local folders once, allowing users to search by memory and open the best match without the need for extensive archive digging. This tool is ideal for freelancers, small engineering teams, and prototype shops that need to recover repeat work across messy client archives. By making archive knowledge searchable and accessible, CADCrawl helps teams stop relying on memory alone and ensures that valuable historical work is not lost or duplicated.