Baidu releases Unlimited-OCR under permissive MIT license
The Chinese tech giant's multilingual vision-language model targets text extraction across languages and document types.

Baidu has published Unlimited-OCR on Hugging Face, a multilingual vision-language model aimed at optical character recognition. The release lands under the MIT license, one of the most permissive options available, meaning developers can use, modify, and ship the model commercially with minimal restrictions.
Unlimited-OCR is built as a vision-language system rather than a traditional OCR pipeline. That approach has become increasingly common for document understanding, where models read images and produce structured or plain text while drawing on broader language reasoning to handle layout, context, and multiple scripts.
Why it matters
OCR remains one of the most practical and widely deployed AI tasks, underpinning everything from document digitization to data entry and accessibility tools. A few points stand out about this release:
- The MIT license lowers the barrier for commercial adoption and downstream fine-tuning.
- A multilingual focus suggests the model is intended to work across scripts rather than English-only text.
- Backing from Baidu, a major player in Chinese-language AI, adds weight to the multilingual claim.
Baidu has not yet detailed parameter counts, context length, or benchmark results in the available record, so independent evaluation will be the real test of how Unlimited-OCR compares to established open OCR systems. For now, the model is available to download and try directly from its Hugging Face repository.
Sources
- Visit
baidu/Unlimited-OCR
Hugging Face
More from Baidu
All Baidu releases →
Baidu releases Unlimited-OCR 2 under MIT license
The multilingual vision-language model targets document text recognition and ships with a permissive license for commercial use.

Baidu releases ERNIE 5.5 Flash Base, a lean MoE model
The bilingual mixture-of-experts base model arrives on Hugging Face, extending Baidu's ERNIE family toward faster, lighter deployments.
Baidu's PP-OCRv6 packs 50-language OCR into tiny models
The latest release of PaddlePaddle's optical character recognition suite spans models from 1.5M to 34.5M parameters under an Apache 2.0 license.
More in Vision-Language
All Vision-Language →
Cactus Compute's Mirage brings vision tasks on-device
The new open vision-language model targets detection, segmentation, and visual Q&A without relying on the cloud.

Thinking Machines ships Inkling-2-Small, an open MoE VLM
The Apache-2.0 vision-language model uses a mixture-of-experts design and lands openly on Hugging Face.

Meituan Releases LongCat 2.5, an MIT-Licensed MoE
The Chinese delivery giant expands its open-model push with a multimodal, agentic mixture-of-experts system.
0 comments
No comments yet. Be the first to weigh in.