EcoSync

Inspiration

I set out to build EcoSync because home-energy monitoring today feels scattered and opaque. I craved a single dashboard to visualize, automate and optimize every smart device—so people save money and carbon without a PhD in engineering.

What it does

EcoSync brings together real-time feeds from your smart plugs, thermostats and solar inverter to:

Plot live + historical power use by device
Surface AI-driven tips (“Your coffee maker idled 15 min after brew.”)
Let you craft Automation Rules and Scenes (“When solar hits 80% output, charge EV.”)
Answer natural-language voice commands (“EcoSync, shut off all basement loads.”)

How we built it

Front-end: React + TailwindCSS, with websockets for live updates
Back-end: Node.js/Express, SQLite for telemetry, MQTT broker for device streams
AI layer: OpenAI GPT-3.5 via carefully batched prompts to stay under token budgets
Automation engine: Rule-eval loop in JavaScript, running on a lightweight server

Challenges we ran into

Total deletion: On submission day, the hackathon platform wiped my draft—every line of code, every rule. I had no backup on that host, so I had to rebuild core modules from memory under extreme time pressure.
Solo & token-starved: With no teammates, I juggled frontend, backend and AI-prompt design alone—while rationing API calls to avoid mid-demo cutoff.
Reliability under load: Re-implementing MQTT + websocket sync twice taught me hard lessons about dropped packets and reconnection logic.

Accomplishments that we’re proud of

Rebuilt MVP in 48 hours after total data loss, complete with AI insights and automation rules
Token-saving feats: Slashed AI costs 70% via prompt-chunking, response caching and smart batching
UX milestone: Delivered a polished dashboard in time for the demo, despite the platform meltdown

What we learned

Back up relentlessly: No hackathon submission is complete without off-site version control
Embrace the pivot: Adversity forced me to prioritize core features; nice-to-haves became tomorrow’s roadmap
Solo developer trade-offs: Rapid iteration comes at the expense of peer review—build in extra debug time

What’s next for EcoSync

Add Git-backed auto-save so no more “total wipe” nightmares
Enable multi-user accounts for shared homes or offices
Integrate Zigbee/Z-Wave hubs and inverter vendor APIs
Ship mobile apps with push alerts on anomalies
Open-source core modules and invite the community to extend and harden the platform