Features

Everything InlinePDF does that generic OCR tools don't.

A review editor built around the original page

Every recognized line sits beside the source image. Edit, validate, or accept — one line at a time.

Image and text side by side, desktop and mobile.
Per-line confidence so you know where to look first.
Find & replace across a whole document, with diacritics-aware matching.
Keyboard-first: move, edit, accept, and change pages without touching the mouse.

Arabic-first OCR, not English-first with Arabic bolted on

Diacritics, ligatures, historical fonts, and proper names — recognized with care.

Preserves tashkeel (diacritics) when present.
Handles Arabic-Indic and Western numerals, your choice.
Recognizes headings, paragraphs, tables, footnotes, lists, and figures.
Optional second-model verification flags confident mistakes.

Structure that survives the export

Tables stay tables. Footnotes stay footnotes. Page numbers stay where they belong.

18 line types enforced through the schema, not guessed.
DOCX export includes real tables and centered page numbers.
Markdown, HTML, TXT, and JSON all preserve structure.
Failed pages get clear placeholders, not silent gaps.

Export to the format you actually need

Word, text, Markdown, HTML, or JSON — pick per document, export by page range.

DOCX for editors and reviewers.
Markdown and HTML for the web.
JSON for pipelines and downstream apps.
Export the whole document or a single chapter.