Dwanith
A persistent Windows workspace prototype that watches the desktop, remembers context, and turns files, notes, slides, clipboard contents, downloads, notifications, and project state into computational objects.
Core trick
Captures Windows screens on a loop and sends each frame to Gemini for analysis, then flags conflicts against stored project context, citing the frame.
Description
A persistent Windows workspace prototype that watches the desktop, remembers context, and turns files, notes, slides, clipboard contents, downloads, notifications, and project state into computational objects. It mainly work by continuous Windows screen capture with Gemini Interactions analysis/through Gemma local continuous screen observation.
For example, if you make some changes, lets say to your codebase or video edit or presentation slides that are incompatible, Persistent Computer will respond: This action conflicts with what I know about your project. Here is the evidence, and here are reversible options.
Normally what we do is we interact with software. But with ai, we can make software interact with us such that every project can be sort of a living software. Just like tom riddle's dairy, I have built persistent computer to notice any issues, flaws or etc in the system and point them out to the user.
dwan-ith/persistent-computer★ 00 forksJavaScriptApache-2.0Created Jul 11, 2026 · last push Aug 30, 2026
README excerpt
# Persistent Computer A persistent Windows workspace runtime that continuously observes the desktop, builds context from workspace files, and intervenes when it detects meaningful visual or structural anomalies. The goal is not another dashboard. The goal is a local computer layer that can say: > This action conflicts with what I know about your project. Here is the evidence, and here are reversible options. ## What It Does - Continuously captures the Windows primary screen and runs Gemini multimodal analysis every cycle - Detects visual breakdowns, contextual conflicts, and structural issues with grounded evidence - Pushes intervention cards into a live sidebar with screenshot-grounded Nano Banana images - Maintains a deduplicated issue history so the same problem is never surfaced twice - Stores persistent memory: facts, episodes, and cross-surface associations used in every Gemini interaction - Provides a native companion arrow that follows the real cursor outside the browser window - Exposes workspace agents for notes, gallery, files, downloads, dependency graph, clipboard, and notifications - Supports multilingual voice output via Gemini-powered translation, with natural TTS playback - Accepts microphone input routed through the Gemini API for conversational interaction with the desktop agent ## Architecture ``` Windows primary screen -> PowerShell screen capture (PrimaryScreen bounds) -> Node observer daemon -> Gemini 3.5 Flash reasoning / Nano Banana image