Inspiration

In today's fast-paced digital workflow, constantly switching between windows, menus, and physical inputs can introduce friction and fatigue. More importantly, for individuals with limited mobility or motor impairments, standard mouse-and-keyboard navigation presents severe accessibility barriers. We built SentinelAI to create a lightweight, responsive bridge between intuitive web/voice interfaces and host desktop operations.

What it does

SentinelAI functions as an assistive local desktop control and monitoring suite:

Real-time Telemetry: Streams live CPU load, RAM utilization, and system specs via low-latency API polling.
Action Interface: Allows one-click or text/voice execution of system tasks, native tools (like Notepad, Calc), and browser routines.
Hands-Free Capability: Structured to ingest natural commands, removing dependency on continuous manual peripheral inputs.

How we built it

Frontend:** Modern, responsive dark-themed dashboard crafted with semantic HTML5, modular CSS3, and vanilla JavaScript.
Backend:** Local lightweight Python REST server exposing dedicated /api/metrics and /api/action endpoints interfacing directly with OS-level subroutines.
Integration:** Low-overhead polling architecture that maintains synchronized hardware diagnostics with the web client.

Challenges we ran into

Ensuring local OS task execution without blocking the main telemetry polling thread.
Managing safe cross-origin execution between browser-initiated triggers and local system processes.
Coordinating rapid prototyping, frontend-backend handshakes, and deployment within a strict 48-hour sprint.

Accomplishments that we're proud of

Successfully delivering an end-to-end functional prototype bridging web interfaces directly to local desktop automation.
Engineering a clean, zero-bloat dashboard that runs smoothly without hogging background system resources.
Building an adaptable, modular backend base ready for future autonomous agent layers.

What's next for SentinelAI

Localized speech-to-text integration for full hands-free voice commanding.
Expanding LLM-driven automation agents capable of chaining multi-step desktop workflows.
Cross-platform support across macOS and Linux distributions.