deepseek-ocr-client

Why we included this project

DeepSeek-OCR does solid document extraction, but the model normally expects Python and a terminal, which is a hurdle if you just want to scan a screenshot. This client wraps it in an Electron desktop app: drag an image onto the window, hit Run OCR, and the text comes back in real time. You can click a region to copy just that part, or export the whole result as a ZIP with markdown images. The app handles model download and loading itself, and it uses CUDA or Apple Silicon GPUs when available, falling back to CPU otherwise. Windows is the best-supported platform; Linux and macOS are still experimental, so keep that in mind before you commit.