Arabic OCR & Verbatim Translation

I have high-resolution scans of Ibn Abī Jamrah’s “Bahjat al-Nufūs” that need to be turned into fully searchable, error-free text and then translated word-for-word into English. Your first task is to run OCR on every page, but you cannot rely on the software alone: each line must be manually compared to the original image so the resulting Arabic in plain-text (.txt) form is absolutely complete and accurate—no missing phrases, no garbled diacritics.

Once the Arabic draft is locked, feed that exact text to ChatGPT and extract a literal, unsummarised English rendering. I do not want an edited or polished version—just the faithful output. Afterward, read through the English side while cross-checking with the Arabic source to confirm that nothing was shortened, merged, or omitted by the model.

File structure I need:

• One plain-text file per Arabic section or chapter
• A separate plain-text file holding the corresponding English for each section

Deliver both sets in clearly numbered folders so I can line them up at a glance.

Acceptance criteria: every word present, in order, with no summaries; Arabic text passes manual comparison to the scans; English text aligns one-to-one with the Arabic source.

No correction of the translation is necessary as long as it is complete and unsummarised.

https://usul.ai/t/bahjat-al-nufus-wa-tahliha-bi-marifat-ma-laha-wa-alayha