What it does
Umi-OCR is an open-source recognition tool for screenshots, images, and PDFs. It can extract scanned-PDF text or create searchable PDFs, with layout processing and excluded recognition regions.
Choose document OCR for scanned PDFs or extract existing text, then save a searchable PDF or text output. Set the column/line-processing mode for the source layout. To suppress repeated headers or watermarks, draw ignore regions large enough to enclose each whole OCR text block; partial characters alone are not discarded. The free MIT-licensed app runs offline on Windows and Linux. Its local HTTP API supports task polling and PDF, TXT, JSONL or CSV results, with ZIP packaging for multiple output types. Retrieve results before clearing the task, which invalidates its download links.
Strengths
- Provides recognition exclusion regions and layout post-processing.
Watch out for
- Ignore regions act on whole OCR text blocks; engine-specific parameters and document-API version requirements matter.
Best for
- Batch recognition of scanned PDFs with repeated page furniture