OCR & document parsingFree

LightOnOCR-3

LightOn's Apache-2.0 OCR models (0.8B, 1B and 4B) that transcribe pages and can also return layout boxes and chart tables.

Official site: huggingface.co ↗

Is it actually free? Free

Free under Apache 2.0. You run it on your own hardware (LightOn's speed figures are for an H100 GPU); a free demo runs on Hugging Face.

Pricing

Free, open weights under Apache 2.0.

Best for

Turning scanned PDFs into text plus labeled layout boxes and chart data.

Not great for

Quick jobs on a laptop with no GPU.

Alternatives

OCR & document parsingFree

Indra OCR

Datalab's OCR model that turns scans and PDFs into markdown on a plain CPU, first published Oct 10–13.

Free? Free to download and run on your own computer. Code Apache-2.0. Weights: modified OpenRAIL-M, free for personal use, research, and organizations at or…

OCR & document parsingFree

Docling

Open-source toolkit that converts PDFs, Office files and images into structured markdown or JSON for AI pipelines. v2.139.0 released Oct 14.

Free? Free under the MIT license. You install it with pip and run it locally or in your own pipeline.

Last checked 2026-10-15 · Source: huggingface.co pricing ↗ · huggingface.co pricing ↗ · How we verify · Report an error