Supported documents and file types

The platform accepts PDF, DOCX, XLSX, CSV, and plain text. Scanned PDFs work but produce weaker results than native text PDFs, because character recognition introduces errors that propagate into the analysis. If you have a native version of a document, use it. Very large files are processed in segments, so a long CIM may take longer to appear than a short teaser. If a document has been processing for more than a few minutes, contact us with the deal reference.