A converter that tells you what changed.

FileType Converters turns any of 65 file formats into any other — 584 conversions — and is honest about the result. Every conversion reports what was kept, what was dropped, and where it ran.

Who we are

FileType Converters is built around one promise: make file conversion transparent, specific, and useful. We have spent years moving files between systems that disagree about what a “document” or a “table” is, and wanted a converter that treats those disagreements as information rather than hiding them.

The product is the converter. The 584 conversion pages exist so that someone searching for “XLSX to CSV” lands on a working tool, a specific list of what that conversion loses, a worked example, and the equivalent command — not a slogan.

How files are converted

Table and data engine: CSV, TSV, JSON, YAML, XML, XLSX, ODS, Parquet, SQL, SRT, VTT
Our own parser and writers, so warnings name the exact cell, sheet or key that changed. Every conversion counts toward your allowance.
Document engine and PDF renderer: Markdown, HTML, DOCX, ODT, RTF, reStructuredText, LaTeX, EPUB → each other and PDF
Document structure (headings, lists, tables, footnotes) is preserved; PDFs are typeset fresh from that structure.
Office engine: DOC/DOCX/ODT, XLS/XLSX/ODS, PPT/PPTX/ODP → PDF, HTML and each other
Used when page-level fidelity matters more than structure.
Ebook engine: EPUB, MOBI, AZW3, FB2
Ebook containers, metadata and reading order.
Image engine: PNG, JPG, WebP, GIF, BMP, TIFF, AVIF, HEIC, ICO, SVG, PSD
Colour-managed resizes and re-encodes with metadata stripped by default.
AI Deep Read: Scanned PDFs, photos and screenshots → Markdown, text, JSON, DOCX, HTML
A vision model reads page images. Pages are sent for reading only and are never used for training.

Your files

Retention: Files are encrypted in transit, processed in a temporary volume, and the input is deleted within 24 hours. Outputs follow your workspace retention setting.
Training: We never use your files to train models. AI Deep Read sends page images to the vision model for reading only.
Access: Workspace members with the right role can re-download outputs during the retention window. Support staff cannot open your files.
Sub-processors: Hosting, payments (Dodo Payments), email (Resend) and analytics (PostHog) are listed in the privacy policy.