Scanned PDF to Markdown — OCR + Conversion in One Step
A "scanned" PDF is just images of pages — there's no text underneath. Most PDF tools stare at it blankly. Our converter detects when a PDF has no text layer, runs OCR automatically, and emits Markdown the same way as for digital sources. One upload, one download, no separate OCR step.
How we detect a scanned PDF
The first heuristic is text density: if the PDF's text layer covers less than 10% of typical page area, we treat it as scanned. The second is glyph quality: fonts that look like the PDF was OCR'd badly already (mojibake, garbled spacing) get re-OCR'd. Either way, the user sees the same flow — drop the file, get Markdown back.
What OCR can and can't recover
Cleanly-scanned typed text at 300+ DPI: near-perfect, often above 99% accuracy. Lower-resolution scans (faxed contracts, photos taken on a phone): 90–98% accuracy with occasional confused pairs (l/1, O/0). Handwriting: variable — neat block printing works, cursive notes are a coin-flip. Tables in scanned form: rows and columns recovered, but cell alignment is best-effort.
Before / After
Before (PDF):
[Scanned image of a page]
(no extractable text — all glyphs are pixels)
After (Markdown):
# Service Agreement
This Agreement is entered into as of January 15, 2026, by and between Acme Corp ("Provider") and Beta LLC ("Client").
## 1. Scope of Services
Provider shall deliver...
Frequently asked questions
How does OCR handle scanned PDFs in this tool?
Automatically. We detect the absence of an extractable text layer (or a corrupted one) and run OCR before conversion. The output is regular Markdown — you don't have to know it was a scan.
What's the accuracy on low-quality scans?
There is no guaranteed accuracy percentage. Low resolution, skew, poor lighting and complex layouts can cause missing or incorrect text. Compare the result with the original scan, especially names, numbers and table cells.
Can I OCR handwritten PDFs?
Block printing and clean cursive: usable, with manual review for confused characters. Doctor-style or rapid notes: not reliable enough to trust without careful proofreading. The OCR engine flags low-confidence regions in the output for you to inspect.
Does OCR slow down conversion significantly?
Scanned-PDF conversion takes ~3–10× longer than digital-PDF conversion of the same page count, because OCR is the dominant cost. For a 50-page scan, expect 15–60 seconds depending on quality.
Are scanned tables converted correctly?
Row and column structure is recovered for most clearly-bordered tables. Borderless tables, merged cells, or tables with rotated text are best-effort and often need manual cleanup. The converter outputs GFM tables when confidence is high and falls back to plain text rows otherwise.