Voltnest - energy predictor

• 

  demo version of a quick simulation

• 

• 

• 

  ability to add appliances

• 

  Shows the predicted energy usage

• 

  gives a room using an AI agent according to the english explaination

• 

• 

  suggests MCBs for the designed circuit in the predictor

• 

  Still not fully functional due to the time limit of coding.Supposed to suggest the ways to reduce the energy usage

Inspiration

Most people have no idea how their home's wiring or electricity use works until the bill arrives or a breaker trips. Electrical planning is full of jargon, so we wanted a tool anyone can use with zero electrical knowledge.

What it does

Voltnest lets you design a home's electrical setup and see how it behaves:

• Describe your house in plain English, or build it room by room, and an AI agent turns it into a model you can review
• Simulates a typical day, month and year with realistic habits (lights at night, average run times for washing machines, vacuums and more)
• Estimates energy use in kWh (and cost), using editable default wattages when you don't know a device's power
• Suggests MCB (safety switch) ratings per circuit and shows a friendly warning when a circuit overloads
• An "Improve my home" advisor agent proposes fixes, re-simulates and shows before/after savings
• Cartoon icons instead of electrical symbols, plus a diagram export [delete if not built]

How we built it

• Written in Jac, using [graph/walkers, byLLM, full-stack Jac: list only what you used], hosted on **Jac Hammer
• A deterministic simulation engine does all the maths, for example $E = P \times t / 1000$ (kWh) and $I = P / 230$ for circuit current
• AI agents handle parsing and advice only; they never invent the numbers
• Built using [GPT-6 Luna / model you used] for the agents, with a fallback to sample houses if the AI call fails

Challenges we ran into

• Keeping the project at least 80% Jac while using AI to write code
• Making realistic default usage patterns for appliances
• Explaining technical results in plain language
• Building everything within a single day

Accomplishments that we're proud of

• A working end-to-end app: describe a house, simulate it, get advice
• Agents that simulate, find the problem, suggest a fix and re-simulate
• An interface a non-technical person can use without help

What we learned

• How to build with Jac and Jaseci [add specifics]
• Working with AI in small, testable steps
• Keeping AI for reasoning and a deterministic engine for numbers

What's next for Voltnest - energy predictor

• Forecast future bills from real usage and local tariff plans
• Connect to smart-meter data
• Solar and battery what-if planning
• Country-specific wiring standards and drag-and-drop floor plans
• A shareable report to hand to an electrician

Built With

• claude
• gemini
• jac
• jachammer