Umi-OCR
Extract text from image batches while excluding repeated watermark or header regions.
- Starting price
- Free local software
- Free access
- No hosted usage allowance promised
- Billing unit
- Local runtime; own hardware
- Platform
- Windows portable; Linux build route
- Verified
- 21 Oct 2026
Overview
What you provide
- Drag images into a batch tab and define OCR language, engine and optional ignored regions.
What you get
- Umi-OCR produces recognized text with configurable post-processing and batch output options.
Choosing this tool
Umi-OCR fits a batch of screenshots or scanned images where recurring headers interfere with text extraction. Its ignore-region editor gives a concrete reason to choose it over a single-screen capture tool. NormCap is lighter for one text snippet; OCRmyPDF is better when preserving a searchable PDF is the primary deliverable.
A decision worked through
For 200 screenshots with the same footer, configure the ignored rectangle using samples from different image sizes. Excluding a region can also erase wanted text if the layout shifts. Inspect several outputs before applying the batch rule; a free run can still create expensive cleanup work.
Umi-OCR alternatives for batch image ocr with ignored regions
For Umi-OCR, compare NormCap, OCRmyPDF for batch image ocr with ignored regions. Check the delivery format and operating responsibility as well as price.
| Product | Price / unit | Free allowance | Input → output | Key limit |
|---|---|---|---|---|
| Umi-OCRRead profileOfficial source ↗ | Free MIT desktop OCR | Offline engines; no page pack | Image batches become OCR text. | Ignore regions can omit real text. |
| NormCapRead profileOfficial source ↗ | Open-source capture app | No per-capture software fee | Screen region becomes usable text. | Screen permissions and language data. |
| OCRmyPDFRead profileOfficial source ↗ | Free local PDF pipeline | No per-page service quota | Scanned PDF becomes searchable PDF. | Tesseract and Ghostscript needed. |
Setup & workflow
- Choose a runtime package and test one image before adding plugins.
- Set ignored regions only after checking several representative layouts.
- Review batch output and retain the original images for corrections.
Requirements & installation
- Extract the official Windows package or follow the separate Linux runtime build instructions; choose one initial engine.
Limits & review
- Do not install both Scoop variants blindly because shortcuts can overwrite one another. Region rules are coordinate-sensitive, and alternate engines may produce different reading order or recognition results.
Your part
- Inspect skipped regions, reading order and numbers on multiple layouts before accepting the batch.
Sources & verification4
This profile is based on official sources, not a hands-on product test.
Vendor descriptions and demos document advertised features. Editorial guidance is based on these sources.
- Official MIT licensegithub.com
- Official Umi-OCR batch workflowgithub.com
- NormCap official referencegithub.com
- OCRmyPDF official referencegithub.com
