Versions:
VideOCR is an open-source utility developed by timminator that focuses on the extraction of hard-coded (burned-in) subtitles from any video file through a lightweight graphical interface. Instead of re-encoding or cropping, the program captures the on-screen text by applying the PaddleOCR machine-learning engine frame-by-frame, turning visible lines into editable subtitle files that can later be translated, indexed or archived. The entire workflow is controlled from a single window: users load a clip, set the language model, optionally limit the detection zone to the lower part of the picture, and start the recognition task; progress, confidence scores and a side-by-side preview are updated in real time, while advanced settings let experienced operators adjust OCR sensitivity, frame sampling rate and output format (SRT, VTT or plain TXT). Typical scenarios include digitising legacy TV recordings that contain no separate subtitle track, creating transcripts of lectures or documentaries for accessibility compliance, and localising fan-made releases where text is permanently merged into the image. The software belongs to the “Video – Subtitle & OCR” category, runs on 64-bit Windows 7 or newer, and is delivered as a portable executable so no system-wide installation is required. Only one build has been published so far, version 1.3.0, released under the MIT licence on GitHub, with all dependencies such as PaddleOCR, FFmpeg and Tkinter bundled inside a 120 MB package. Because recognition happens locally, no Internet connection or cloud quota is consumed, and extracted subtitles remain private. VideOCR is available for free on get.nero.com, where downloads are served through trusted Windows package sources (e.g., winget), always supply the latest version, and support batch installation of multiple applications.
Tags: