Umi-OCR

Umi-OCR extracts text from screenshots and local images for people who need to copy on-screen text or process image collections without an online OCR service.

• Screenshot text: Recognize a captured screen area or pasted image, then select, copy or edit the extracted text.
• Batch recognition: Process local images together and export results as plain text, Markdown, JSONL or CSV.
• Text layout: Merge lines into paragraphs or preserve spacing in code snippets; vertical text requires a compatible recognition model.
• Ignore regions: Exclude areas containing recurring watermarks or logos from batch results.
• QR codes and barcodes: Read codes from screenshots and images, including multiple codes in one image, or generate code images from text.

Umi-OCR is free and open source under the MIT license. It supports 64-bit Windows 7 and later and includes an offline OCR engine and language libraries, so recognition doesn't need a network connection. Rapid-OCR and Paddle-OCR are available as engine choices through plugins. The interface supports languages including English, Japanese and Traditional Chinese, and command-line and HTTP interfaces let other software call its OCR functions.