Working demo · built for your Upwork brief

EV charging across every station, managed by AI.

Voltara is a working prototype of the app you described: drivers register with OTP, find a free charger, start and stop sessions, and watch live charging status. AI plans cheaper charging and balances power between stations in real time.

Real backend on AWSKimi K2.5 via Amazon Bedrock9 stations · 30 chargers liveNo login needed to explore
Live app · tap around, it’s real
Three things to try

Two minutes, three moments.

Every one of these runs end to end on the live backend. Nothing is a screenshot.

1

Start a live session

Pick the AI-suggested charger, hold to start, and watch a full CC-CV charging curve, cost and range tick up. 60× time-lapse makes a 40-minute charge take about 40 seconds. You get a GST receipt at the end.

Open the app
2

Throttle it from Ops

With a session running, open the operator console in another tab or phone and drag that site’s grid cap down. Dynamic load balancing re-shares power and the driver sees the drop, with the reason, within seconds.

Operator console
3

Ask the AI

“Cheapest way to 80% by 8 am?” The optimiser picks 15-minute tariff slots and the AI explains the plan. The assistant also answers trip and charger questions from live network data, in English, Hindi or Gujarati.

Ask Voltara AI
Mapped to your brief

Everything in the job post, working.

User registration

Phone + OTP sign-up, vehicle profile, wallet and history. Guest mode for instant try-out.

“support user registration”

Real-time charging status

Live SoC, kW, kWh, cost, ETA and a charging curve, updated every ~1 s from the server.

“real-time charging status updates”

Manage sessions

Choose a target %, hold-to-start and hold-to-stop, receipts with GST, wallet deduction.

“managing charging sessions”

Multiple stations

9 Rajkot sites, 30 chargers (AC, CCS2, Bharat DC-001) with live availability and busy-hours.

“across multiple stations”

AI load balancing

Each site shares its grid cap across cars, lowest battery first. Operators change caps live.

“uses AI to help manage EV charging”

AI assistant & smart charging

Tariff-aware schedules, savings vs “charge now”, and a chat assistant grounded in live data.

“simple, intuitive interface”
Honest about the demo

What’s real, and what’s simulated.

Real, running on AWS

  • OTP sign-up, signed sessions, wallet and receipts (DynamoDB)
  • Server-side charging sessions: battery physics, CC-CV taper, per-site load balancing
  • Operator controls (grid cap, balancing mode, remote reset) that affect live sessions
  • Smart-charge optimiser over 15-minute tariff and carbon slots
  • LLM answers streamed from Kimi K2.5 on Amazon Bedrock through an IAM role, with no API keys

Simulated for the demo

  • Charger hardware. Production connects over OCPP 1.6J / 2.0.1 (start, stop, meter values, reset)
  • SMS delivery (the OTP is shown on screen), UPI / Razorpay top-ups, GPS position
  • Background traffic from other drivers (a deterministic demand model) and the tariff values
  • Time-lapse speeds (1× / 20× / 60×) so you can see a full session quickly