OCRmyPDF vs olmOCR
Pick OCRmyPDF to make scanned PDFs searchable and keep them as PDF/A on a CPU-only machine. Pick olmOCR to turn PDFs and images into Markdown or JSON for RAG or LLM training, if you have an NVIDIA GPU with at least 12 GB VRAM. The main split is the output: OCRmyPDF adds a text layer to the original file, while olmOCR rewrites the content as structured text.
OCRmyPDFocrmypdf/ocrmypdf | olmOCRallenai/olmocr | |
|---|---|---|
| Type | Open source | Open source |
| Price | 1 credit / page | 2 credits / page |
| Per 1,000 pages | $1.00 | $2.00 |
| Typical speed | 1.0 s / page | 1.5 s / page |
| Best for | Searchable PDFs from scans | Bulk PDF to text for datasets |
| Inputs | PDF, images | PDF, images |
| Languages | 100+ languages (via Tesseract) | Mainly English |
| Max pages per request | 1,000 | 1,000 |
| Tables | No | ✓ Yes |
| Handwriting | No | ✓ Yes |
| Math | No | ✓ Yes |
| Bounding boxes | No | No |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | ✓ Yes | No |
| On OCRRouter | Available | Available |
Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.
Which to choose
Choose OCRmyPDF if
- You need validated searchable PDF/A output for archiving, with other PDF content left unre-encoded where possible.
- You only have CPUs, and you want work spread across all cores.
- You need Tesseract's 100+ languages through language packs.
Choose olmOCR if
- You need Markdown or JSON output for RAG or LLM training, not a searchable PDF.
- Your documents contain equations, tables or handwriting, and you want headers and footers stripped in reading order.
- You process pages at scale and want self-hosting under $200 per million pages, per Ai2, or a hosted endpoint at $0.09 input / $0.19 output per million tokens on DeepInfra.
Key differences
| OCRmyPDF | olmOCR | |
|---|---|---|
| Output | Searchable PDF (PDF/A by default) and plain text | Markdown and JSON |
| Hardware | CPU only, uses all cores | NVIDIA GPU with at least 12 GB VRAM and 30 GB free disk |
| Licence | MPL-2.0 | Apache-2.0 (toolkit) |
| Price | Free, self-hosted (you pay for compute) | Self-hosted: under $200 per million pages, per Ai2. Hosted olmOCR-2-7B on DeepInfra: $0.09 input / $0.19 output per million tokens |
| Layout, tables, handwriting | No layout, table or Markdown output; not suited to handwriting | Handles equations, tables and handwriting |
| Benchmark | Not published | 82.4 on olmOCR-Bench (v0.4.0), reported by the authors |
| Known failure | --skip-text with default PDF/A output removes form fields (issue #1790); --deskew --clean-final can double or quadruple file size (issue #948) | Out of memory on a 12 GB GPU (issue #424); tends to get stuck on the last few pages (issue #433) |
Sources
- OCRmyPDF GitHub README
- Issue #948 file size
- Issue #1790 form fields
- olmOCR GitHub README
- Issue #424 OOM
- Issue #433 stuck
What each one is
OCRmyPDF
Command-line tool that adds a searchable text layer to scanned PDFs, using Tesseract.
OCRmyPDF detailsolmOCR
Open-source vision model from Ai2 (often searched as olm ocr) that turns PDFs into clean, reading-order text. Latest release: olmOCR 2.
olmOCR detailsTry both on your own file
Run the file once with each model and compare the output.
Use one, fall back to the other
{"model": "ocrmypdf/ocrmypdf","fallbacks": ["allenai/olmocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
OCRmyPDF alternativesolmOCR alternativesBest OCR models in 2026