Back to browse
GitHub Repository

AI-powered screen memory — captures, analyzes, and lets you search/chat your screen history. Powered by Gemma 4 . 100% local, 100% private.

77 starsPython

I run a vision model on every screenshot, locally, on a 4GB GPU

by skye0110·Jun 13, 2026·36 points·7 comments

AI Analysis

●●SolidDark HorseSolve My Problem

Runs multimodal screen memory on 4GB VRAM while Microsoft Recall requires high-end NPU.

Strengths
  • Content-change detection reduces storage overhead compared to fixed-interval screen recorders.
  • MCP integration lets AI agents query screen history directly inside Cursor.
  • Hybrid search combines semantic embeddings with FTS5 for accurate context retrieval.
Weaknesses
  • Gemma 4 availability is uncertain; relies on unreleased or specific quantized weights.
  • Lacks performance benchmarks to verify the 4GB VRAM claim.
Category
Target Audience

Privacy-conscious developers, knowledge workers, local-first AI enthusiasts

Similar To

Rewind.ai · Microsoft Recall · OpenRecall

Similar Projects