KatavinaNguyen/screenshot_based_ai_desktop_assistant
A lightweight Python-based desktop assistant that lets users capture a region of their screen, extract text using PaddleOCR, and instantly query selected large language models (LLMs) for responses, all without interrupting workflow. Designed with a minimal popup UI and global hotkey support for distraction-free productivity.
Platform-specific configuration:
{
"mcpServers": {
"screenshot_based_ai_desktop_assistant": {
"command": "npx",
"args": [
"-y",
"screenshot_based_ai_desktop_assistant"
]
}
}
}Add the config above to .claude/settings.json under the mcpServers key.
Loading reviews...