# PaddleOCR

19 Hacker News commenters recommended PaddleOCR and 4 criticized it, November 2025 to November 2026. It was recommended in 8 of 13 months. It meets the Hacker Approved badge rule.

**Verdict:** Free open-source OCR toolkit that people use to extract text and layout from documents, running fast on CPU. Critics say it trails some alternatives on accuracy.

## Why people recommend it

- Accurate text and layout extraction (7 comments): "I just built a pipeline for document data extraction that uses PaddleOCR, then Gemini 3 to check + fix errors. It gets close to 99.9% on extraction from financial statements" ([source](https://news.ycombinator.com/item?id=46332411)) "it correctly split the text columns, labled the illustrations, and noted the maragin text" ([source](https://news.ycombinator.com/item?id=47853350)) "Paddleocr on single line text for me has accuracy in the range of 95%" ([source](https://news.ycombinator.com/item?id=49956215))
- Fast and runs offline on CPU (4 comments): "it's definitely fast enough. Probably faster than the Sonnet call." ([source](https://news.ycombinator.com/item?id=48383572)) "working reasonably well and fast on my CPU laptop" ([source](https://news.ycombinator.com/item?id=48414063)) "Works pretty well. Runs fully offline, in CPU." ([source](https://news.ycombinator.com/item?id=48648011))
- Works in document pipelines (2 comments): "I just built a pipeline for document data extraction that uses PaddleOCR, then Gemini 3 to check + fix errors. It gets close to 99.9% on extraction from financial statements" ([source](https://news.ycombinator.com/item?id=46332411)) "PaddleOCR, Spacy NER, BGE-M3 embeddings." ([source](https://news.ycombinator.com/item?id=46286667))

## What people warn about

- Trails other OCR tools on accuracy (3 comments): "Tasserect is the worst compare to PaddleOCR , Surya/Marker" ([source](https://news.ycombinator.com/item?id=49293634)) "worse performance than many OSS offerings like Paddle" ([source](https://news.ycombinator.com/item?id=46331109)) "performed significantly worse" ([source](https://news.ycombinator.com/item?id=49421088))
- Setup problems on some machines (2 comments): "It took an hour and a half to install 12 gigabytes of pytorch dependencies that can't even run on my device" ([source](https://news.ycombinator.com/item?id=46331521)) "Paddle wasn't working on my local machine" ([source](https://news.ycombinator.com/item?id=46333378))

## Alternatives hackers recommend

- [Mistral OCR](https://hackerapproved.com/p/mistral-ocr.md): 69 recommend
- [GLM-OCR](https://hackerapproved.com/p/glm-ocr.md): 16 recommend
- [DeepSeek-OCR](https://hackerapproved.com/p/deepseek-ocr.md): 12 recommend
- [dots.ocr](https://hackerapproved.com/p/dots-ocr.md): 8 recommend
- [Unlimited OCR](https://hackerapproved.com/p/unlimited-ocr.md): 7 recommend

## Cite as

Hacker Approved, "PaddleOCR", https://hackerapproved.com/p/paddleocr/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).