RapidOCR vs dots.ocr
Pick RapidOCR for fast offline CPU text extraction from Chinese and English screenshots and receipts, or when you need to embed OCR in Java, C# or C++ apps. Pick dots.ocr for multilingual PDFs where you need layout, tables as HTML and formulas as LaTeX in one pass, and you have a CUDA GPU. The main split: RapidOCR returns plain text and boxes, dots.ocr parses whole documents.
RapidOCRrapidai/rapidocr | dots.ocrrednote/dots-ocr | |
|---|---|---|
| Type | Open source | Open source |
| Price | 1 credit / page | 2 credits / page |
| Per 1,000 pages | $1.00 | $2.00 |
| Typical speed | 0.6 s / page | 1.5 s / page |
| Best for | Lightweight OCR on CPU or edge | Multilingual document parsing |
| Inputs | images | PDF, images |
| Languages | Chinese, English and more | 100+ languages |
| Max pages per request | 100 | 500 |
| 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
- It runs on CPU through ONNX Runtime with a plain pip install of rapidocr and onnxruntime, and no Paddle framework.
- It has bindings for Python, C++, Java and C#, so it fits desktop and server apps in many stacks.
- It suits simple Chinese and English screenshots and receipts, where you only need text and bounding boxes.
Choose dots.ocr if
- It returns layout boxes, categories and text in one pass, with tables as HTML and formulas as LaTeX.
- It outputs JSON, Markdown, HTML, LaTeX and bounding boxes, so multilingual PDFs can go straight to Markdown with layout.
- It has published benchmark scores: 79.1 on olmOCR-bench for the original model and 83.9 for dots.mocr (v1.5), both reported by the authors.
Key differences
| RapidOCR | dots.ocr | |
|---|---|---|
| What it is | OCR toolkit that runs PaddleOCR models converted to ONNX and other runtimes | 3B-parameter document parsing VLM (1.7B LLM backbone) |
| Licence | Apache-2.0; converted models keep their upstream licences | MIT; the repository also includes a separate dots.ocr license agreement file |
| How you run it | Local library for Python, C++, Java or C#, on CPU by default (optional GPU, TensorRT, OpenVINO, MNN engines) | Self-hosted via vLLM or Hugging Face Transformers; single CUDA GPU with vLLM is the documented path, no VRAM figure stated |
| Output formats | Plain text and bounding boxes | JSON, Markdown, HTML, LaTeX, bounding boxes and SVG |
| Benchmarks | Not published | olmOCR-bench 79.1 (dots.ocr) and 83.9 (dots.mocr); OmniDocBench v1.5 text edit distance 0.031 for dots.mocr (lower is better); all by the authors |
| Languages | Chinese and English by default; other languages need extra models | Multilingual |
| Known failure | 16-bit grayscale PNG or TIFF give no text or wrong text; sliced inference crashes on large pages; Arabic recognition crashes because python-bidi is undeclared | Runs of ellipses or underscores can make output repeat endlessly; a table above 1000 grid cells raises an error and drops the whole page in Docling; complex tables and formulas remain difficult |
| Price | Free, self-hosted (you pay for compute) | Free, self-hosted (you pay for compute) |
Sources
- RapidOCR GitHub README
- Issue #776
- dots.ocr GitHub README
- dots.ocr blog / limitations
- Docling issue #5146
What each one is
RapidOCR
PaddleOCR models packaged for ONNX Runtime and other engines. Easy to deploy without Paddle.
RapidOCR detailsdots.ocr
Single vision model from rednote that does layout detection and content recognition together.
dots.ocr 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": "rapidai/rapidocr","fallbacks": ["rednote/dots-ocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
RapidOCR alternativesdots.ocr alternativesBest OCR models in 2026