PaddleOCR
Hacker ApprovedRanked 2 of 11 OCR models by recommenders
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.
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.
23 people gave an opinion
- recommend it
- 19
- criticize it
- 4
- approval share
- 83%
Recommended in 8 of 13 months
- November 2025: not recommended
- December 2025: recommended
- January 2026: not recommended
- February 2026: recommended
- March 2026: not recommended
- April 2026: recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: not recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it carries the badge
- 19 people recommend itNeeds 10 or more, met.
- 4.8 recommenders for every criticNeeds 2 to 1 or better, met.
- Recommended in 8 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Accurate text and layout extraction
“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”
Comment from 20 Dec 2025 in “Mistral OCR 3” “it correctly split the text columns, labled the illustrations, and noted the maragin text”
Comment from 21 Apr 2026 in “Britannica11.org – a structured edition of the 1911 Encyclopædia Britannica” “Paddleocr on single line text for me has accuracy in the range of 95%”
Comment from 4 Oct 2026 in “Show HN: AI search for every photo and every frame of video on macOS” - Comment from 14 Aug 2026 in “Mistral OCR 4.1”
- Comment from 8 Jul 2026 in “Decoding the obfuscated bash script on a Uniqlo t-shirt”
- Comment from 13 Feb 2026 in “Gemini 3 Deep Think”
- Comment from 23 Jun 2026 in “Unlimited OCR: One-Shot Long-Horizon Parsing”
Fast and runs offline on CPU
“it's definitely fast enough. Probably faster than the Sonnet call.”
Comment from 3 Jun 2026 in “Show HN: Edsger – A handwritten Clojure REPL for the reMarkable 2” “working reasonably well and fast on my CPU laptop”
Comment from 5 Jun 2026 in “Show HN: Local-first fast CPU image to text for screenshots, PDFs, webpages” “Works pretty well. Runs fully offline, in CPU.”
Comment from 23 Jun 2026 in “Unlimited OCR: One-Shot Long-Horizon Parsing” - Comment from 24 Aug 2026 in “OCR It – pull text out of un-copyable documents for your LLM”
Works in document pipelines
“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”
Comment from 20 Dec 2025 in “Mistral OCR 3” “PaddleOCR, Spacy NER, BGE-M3 embeddings.”
Comment from 16 Dec 2025 in “ArkhamMirror: Airgapped investigation platform with CIA-style hypothesis testing”
What people warn about
Each mark is one comment
Trails other OCR tools on accuracy
“Tasserect is the worst compare to PaddleOCR , Surya/Marker”
Comment from 14 Aug 2026 in “Mistral OCR 4.1” “worse performance than many OSS offerings like Paddle”
Comment from 19 Dec 2025 in “Mistral OCR 3” “performed significantly worse”
Comment from 24 Aug 2026 in “OCR It – pull text out of un-copyable documents for your LLM” Setup problems on some machines
“It took an hour and a half to install 12 gigabytes of pytorch dependencies that can't even run on my device”
Comment from 19 Dec 2025 in “Mistral OCR 3” “Paddle wasn't working on my local machine”
Comment from 20 Dec 2025 in “Mistral OCR 3”
Other OCR models people recommend
The four most recommended other OCR models, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| Mistral OCR | 69 | 40 | 63% | 12 | — | — |
| PaddleOCRThis page | 19 | 4 | 83% | 8 | — | Approved |
| GLM-OCR | 16 | 4 | 80% | 7 | — | Approved |
| DeepSeek-OCR | 12 | 4 | 75% | 9 | — | Approved |
| dots.ocr | 8 | 1 | 89% | 5 | — | — |
What people said
20 of the 24 comments that mention it, newest first. Each links to the original comment.
“Paddleocr on single line text for me has accuracy in the range of 95%”
“most of the chinese open-source libs were license-washing software”
“PaddlePaddle (v6) is fantastic and fast”
“performed significantly worse”
“PaddleOCR is enough to get extremely well-done bounding boxes”
“Tasserect is the worst compare to PaddleOCR , Surya/Marker”
“I ran it through paddle paddle OCR and it flawlessly did it.”
“Works pretty well. Runs fully offline, in CPU.”
“pretty damn good actually”
“for their valuable models and ideas”
“working reasonably well and fast on my CPU laptop”
“it's definitely fast enough. Probably faster than the Sonnet call.”
“if you're looking for a pure CNN, check them out”
“it correctly split the text columns, labled the illustrations, and noted the maragin text”
“it solved most of our issues with ocr”
“Paddle wasn't working on my local machine”
“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”
“It took an hour and a half to install 12 gigabytes of pytorch dependencies that can't even run on my device”
“worse performance than many OSS offerings like Paddle”
“PaddleOCR, Spacy NER, BGE-M3 embeddings.”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.