Baidu Releases Unlimited OCR as Open Source OCR Model for Transcribing Long Documents

Chinese tech giant Baidu has released Unlimited OCR, a 3-billion-parameter optical character recognition model designed to transcribe long documents in a single pass. The company made the tool available under an open-source license, allowing developers and researchers to freely access and modify the architecture. Optical character recognition software converts images of text into machine-readable data, a core component for automating document processing, archival systems, and artificial intelligence retrieval pipelines.

By enabling single-pass processing for lengthy files, the model aims to reduce the need for splitting and stitching documents, which can introduce errors or increase latency in automated workflows. The release positions Baidu to compete with other open-weight vision and transcription models in the developer ecosystem.