An AI energy assistant for Pakistani households. Predict next month's bill, forecast your 24-hour load curve, and simulate appliance swaps — all calibrated to real consumption data and the latest tariff slabs.
Slab jumps, protected-status thresholds, fuel adjustments and time-of-use surcharges make it almost impossible to know what you'll owe until the bill arrives.
Cross a 100 or 200-unit threshold and your entire bill is charged at a much higher per-unit rate — often without warning.
Exceed the 6-month rolling average and you permanently lose subsidised Protected or Lifeline pricing.
Time-of-Use consumers pay up to 25% more for electricity used between 6 PM and 10 PM — easy to miss, costly to ignore.
No smart meter required. Tell us about your home once, and the AI engine handles the rest — recalibrating every time you add a new bill.
Sign up in seconds with email or Google. Your data is synced securely to Cloud Firestore under your account.
Add your DISCO, sanctioned load, property size, and appliance inventory. The more detail, the sharper the prediction.
Random Forest and LSTM models fuse your profile with 42 real Pakistani household archetypes to forecast your bill and load curve.
Swap appliances virtually — standard AC to inverter, old fridge to new — and see the exact Rupee impact before you spend a cent.
Select appliances, adjust their quantities and running hours, and toggle inverter upgrades to see standard vs. efficient energy consumption and estimated savings on your monthly bill.
PKR savings calculated by subtracting efficient load from standard load under NEPRA slabs.
Once your profile is set up, your dashboard fills in with live projections, seasonal alerts, and AI-written insights tailored to your household.
Six connected modules, one shared AI memory of your household.
Random Forest regression estimates next month's units and full NEPRA-calculated cost, including FCA and QTA adjustments.
A Bidirectional LSTM projects your hourly demand curve so you know exactly when your home hits peak load.
A physics-based engine simulates appliance swaps — inverter ACs, BLDC fans — and shows real Rupee savings.
A context-aware chatbot answers questions about your bill using your live profile — no generic advice, just yours.
Every 2026 domestic slab, fixed charge, and Protected/Lifeline eligibility rule, explained in plain language.
One structured household profile — DISCO, routine, appliances, bill history — powering every model in the app.
We combine statistical machine learning, deep learning, and first-principles physics — so predictions stay explainable, not just accurate.
Calibrated on 28+ zero-leakage household features. Handles long-term monthly forecasting by analyzing appliance counts and seasonal drift.
A specialized RNN that recognizes temporal dependencies, processing a 48-hour lookback window to forecast the next 24-hour demand profile.
Maps your profile to the closest match among 42 real PRECON house archetypes, ensuring realistic seeding even without bill history.
Computes a structural baseload from occupancy and area. Final predictions blend physics, AI inference, and historical calibration.
Domestic tariffs in Pakistan aren't a flat rate — they're a maze of slabs, subsidies, and adjustments. Our billing engine models the real structure so predictions match your actual invoice.
Explore Tariff HubPreemptive bill awareness helps households maintain their energy budget before the billing cycle ends.
The LSTM forecaster identifies usage patterns, shifting behaviour toward efficiency and reducing waste.
A four-person final year project turning academic research into a tool real households can use.