What problem does it solve?
It helps users whose current task depends on what’s visible on their screen by providing timely, readable screen context they can reference without asking for repeated manual descriptions.
Core Features & Use Cases
- Launches screen-awareness TUI: Starts the LiteWatch interface that captures periodic screenshots and generates textual summaries for downstream use.
- Reads rolling visual context: Pulls the latest entries from screen_context.txt so the assistant can respond with context aligned to what the user is seeing.
- Supports VLM-backed visual summarization: Uses a locally loaded VLM in LM Studio (and gives guidance if the wrong model type is loaded).
Use this when you need situational awareness for UI troubleshooting, multi-step navigation, or responding to on-screen states like errors, alerts, or newly opened tabs.
Quick Start
Ask for screen awareness by telling the assistant to start litewatch and use it to check what’s on your screen right now.