SolarTrade
ENERGY TECH

SolarTrade

MVP in 38 days

Solar energy traders in India had surplus power but no digital marketplace to sell it peer-to-peer. The existing process was entirely manual — phone calls, spreadsheets, and bank transfers that took 3–5 days to settle.

The client needed a platform that could integrate with DISCOM smart meters in real-time, process UPI payments instantly, and maintain an immutable record of every transaction on a blockchain — all within a tight 38-day MVP timeline.

We built a multi-tenant P2P solar energy trading platform from the ground up. The architecture separates tenant data completely while sharing infrastructure costs — critical for a marketplace with multiple energy producers and buyers.

The DISCOM smart meter integration was the most complex part. We built a real-time data pipeline that polls meter readings, validates energy output, and triggers automated trades when threshold conditions are met.

  1. 1

    FastAPI backend

    Chosen for async performance needed for real-time meter polling.

  2. 2

    PostgreSQL with multi-tenant row-level security

    Keeps producer and buyer data isolated without separate databases.

  3. 3

    Hyperledger Fabric

    Private blockchain for transaction records — not public chain, because the client needed data privacy.

  4. 4

    n8n automation

    Handles settlement workflows and notification triggers without custom code.

  5. 5

    UPI payment integration via Razorpay

    Instant settlement vs 3–5 day bank transfers.

Client

Solar energy trading company, India

Industry

Clean Energy / FinTech

Timeline

38 days (MVP)

Team

Parth + specialist engineers

Status

Live — trading across 3 districts

Tech Stack

FastAPIPostgreSQLHyperledgerUPIDockerReactn8n

Key Results

38 Days to MVP
3 Districts live at launch
100% Blockchain-verified transactions
0

Days to MVP

0

Districts live at launch

0%

Blockchain-verified transactions

DISCOM APIs are inconsistent across states — we had to build a normalisation layer that abstracts the differences. This added 4 days to the timeline but saved weeks of debugging post-launch.