PDF to HTML: iHateUpload vs ihatepdf.cv

Both ihatepdf.cv and iHateUpload run PDF to HTML in your browser without uploading the file. Here is what each one does and what happened when we ran the same synthetic test files through both.

Side by side

• PDF to HTML

  iHateUpload ahead

  iHateUpload

  Three layouts: flowing semantic page (h1-h6 from heading detection, real lists, tables with spans and header rows, links, inline images), exact layout (one positioned page per PDF page: text, tables, pictures and shapes where they were, selectable), or page picture with an invisible text layer; one .html or a zip; OCR for scans; page range; escaped text, lang and dir set; live preview, Copy HTML

  ihatepdf.cv

  Positioned and Semantic modes with themes, per-page export, include toggles and a live preview; no pictures in the output, no OCR; Semantic mode found no table and wrote the numbered list as bullets, and the extractor failed outright on 16 of 19 Chrome-made documents

  How we checked

  Same synthetic files on both (internal/benchmarks/pdf-to-html): on 8 documents they could convert, text 0.99 vs 0.86 (their scan came out empty), headings 1.00 vs 0.89, tables 1.00 vs 0.00, exact visual 0.89 vs 0.63; on a 20-document realistic corpus theirs failed with a script error on 16 (Chrome-made PDFs) while ours converted all 20 with text 0.99, headings 0.87, tables 0.92 (exact layout) to 1.00 (flowing) and a scan read with OCR. Checked 2026-10-21.

Try PDF to HTML, free and without upload