Show HN: ScreenMind – grep for your visual memory, 100% on-device

Category: ai-ml

Tags: ai-memory, screen-capture, local-ai, gemma, privacy, rag

Score: 7.0/10 (Innovation: 6, Technical: 8, Documentation: 7, Utility: 7)

ScreenMind is an open-source, privacy-first alternative to Microsoft Recall that uses the Gemma 4 multimodal model to analyze and index your screen activity entirely on-device. Its innovative use of a single local model for vision, audio, and reasoning tasks, combined with features like hybrid search and an agent platform, makes it a compelling tool for developers and power users seeking a private AI memory for their desktop.

Target audience: backend devs, devops, power users, data engineers

Repository: https://github.com/ayushh0110/ScreenMind · Python · MIT · 22 stars

View on Hacker News