RapidOCR vs Chunkr
Pick RapidOCR for offline Chinese and English text on a CPU, embedded in Python, C++, Java or C# apps. Pick Chunkr when you need layout-aware output (HTML, Markdown, JSON) from PDFs, slides and Word files, for example to chunk documents for RAG. RapidOCR returns only text and boxes, while Chunkr parses document structure.
RapidOCRrapidai/rapidocr | Chunkrchunkr/chunkr | |
|---|---|---|
| Type | Open source | OCR API |
| Price | 1 credit / page | 14 credits / page |
| Per 1,000 pages | $1.00 | $14.00 |
| Typical speed | 0.6 s / page | 3.0 s / page |
| Best for | Lightweight OCR on CPU or edge | Layout-aware chunks for RAG |
| Inputs | images | PDF, images, DOCX |
| Languages | Chinese, English and more | Multilingual |
| Max pages per request | 100 | 1,000 |
| Tables | No | ✓ Yes |
| Handwriting | No | No |
| Math | No | ✓ Yes |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | ✓ Yes | No |
| Forms | No | No |
| Multilingual | ✓ Yes | ✓ Yes |
| 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 RapidOCR if
- You want a library that runs offline on CPU with pip install rapidocr onnxruntime, with no Paddle framework.
- Your app is written in Java, C# or C++, since RapidOCR has bindings for Python, C++, Java and C#.
- Your input is Chinese or English screenshots and receipts, and you only need plain text and bounding boxes.
Choose Chunkr if
- You need HTML, Markdown or JSON output with bounding boxes, not just raw text.
- You feed PDF, PPT, Word or image files into a RAG pipeline and want semantic chunking.
- You want to self-host layout parsing with Docker (GPU, CPU-only or Mac ARM) and can supply an OpenAI-compatible LLM endpoint.
Key differences
| RapidOCR | Chunkr | |
|---|---|---|
| What it does | OCR toolkit that converts PaddleOCR models to ONNX and other runtimes | Document parsing service with layout analysis, OCR and semantic chunking |
| How you run it | Local library for Python, C++, Java or C# | Cloud API, or self-host with Docker (GPU, CPU-only, Mac ARM) |
| Licence | Apache-2.0 (converted models keep their upstream licences) | AGPL-3.0 or commercial licence; closed-source products need the commercial one |
| Price | Free, self-hosted; you pay for compute | Hosted credit-based plans (prices not published); open-source version is free but you pay for compute and an LLM endpoint |
| Output formats | Plain text and bounding boxes | HTML, Markdown, JSON and bounding boxes |
| Input and languages | Chinese and English by default; other languages need extra models | PDF, PPT, Word and images; Excel is not supported in the open-source version |
| Known failure | 16-bit grayscale PNG or TIFF give no text or wrong text; Arabic recognition crashes (python-bidi undeclared) | Open-source version uses community models, less accurate than the cloud's in-house models |
Sources
What each one is
RapidOCR
PaddleOCR models packaged for ONNX Runtime and other engines. Easy to deploy without Paddle.
RapidOCR detailsChunkr
Document parsing API that segments and chunks pages for RAG. Self-hostable version available.
Chunkr detailsTry both on your own file
Run the file once with each model and compare the output.
PDF, PNG, JPG, WebP, HEIC. Up to 20 MB each, 10 files.No file?
Use one, fall back to the other
{"model": "rapidai/rapidocr","fallbacks": ["chunkr/chunkr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
RapidOCR vs ABBYY FineReaderRapidOCR vs UnstructureddocTR vs ChunkrOCRmyPDF vs ChunkrGPT vs ChunkrRapidOCR vs PixtralJina OCR vs ChunkrChunkr vs Adobe Acrobat OCR
RapidOCR alternativesChunkr alternativesBest OCR models in 2026