Private AI — your data never leaves your infrastructure

Your energy data.
Your servers.
Your AI.

Nerqon is a self-hosted Energy Management & BI platform. Connect your databases and machines, see every kilowatt on live dashboards, and ask questions in plain language — answered by an AI that runs entirely on your own hardware.

25+
Data Connectors
66
Prebuilt Analytics
<15 min
To Deploy
0
Cloud Dependencies

Built for industrial teams that own their infrastructure

100% On-Premise
Modbus · OPC-UA · BACnet · MQTT
Local LLM — Zero Cloud
AES-256 Encryption
Air-Gap Ready
BEE & ASHRAE Benchmarks
The Product

Every kilowatt, live.
Every question, answered.

Live dashboards on the left. A private AI that speaks SQL on the right. Both running on your server.

Nerqon — Plant Overview Live
Consumption today
4,832 kWh
6.2% vs last week
Energy cost
₹41,270
today, all meters
Peak demand
412 kW
93% of contract
CO₂ today
3.9 t
on target
Consumption — last 24 hkW
peak 412 kW
00:0006:0012:0018:00now
Ask NerqonLocal AI
Where can we save energy?
Compressor #2 runs loaded during the 23:00–05:00 idle window — about 610 kWh/week above baseline. Rescheduling it to demand-based control would save roughly ₹18,400/month.
machine
idle kWh
₹/mo
Compressor 2
610
18,400
table chart SQL CSV
Ask about your energy data…

Illustration of the Nerqon interface with sample data.

Capabilities

One platform.
Six ways to cut your energy bill.

From live monitoring to AI answers — every capability runs against your own databases and machines, on your own hardware.

How It Works

From bare server to answers
in four steps

One docker compose up -d and you’re running — typically in under 15 minutes.

1

Connect

Point Nerqon at your databases and machines — 25+ connectors from PostgreSQL to Modbus. Every connection performs a real protocol handshake, not just a ping.

2

See

9 prebuilt dashboards light up with your consumption, cost and power quality. Rearrange them or build your own from 15+ widget types.

3

Ask

Question your data in plain language. The local AI answers with numbers, charts and the SQL it ran — grounded in your database, not the internet.

4

Act

Forecast the next 7–90 days, set threshold alerts, track CO₂, and schedule audit-ready PDF/Excel reports for the people who need them.

The Difference

Private AI.
Nothing leaves the building.

Unlike cloud BI tools, the entire Nerqon stack ships as Docker containers on your own server — including a local large language model and a local vector database.

  • No OpenAI keys, no cloud calls, no per-token fees. The AI is trained to answer from your historical readings and schemas.
  • Works fully air-gapped. No internet connection required at runtime — ideal for OT networks and secure sites.
  • Modest hardware. Runs from a single 32 GB server; GPU optional and auto-detected for faster AI answers.
Security & Deployment
Your infrastructure
Docker on your server
Nerqon EMS
Local LLM
14B params
Vector Search
Time-series DB
Report Storage
Connectors
reads your databases, meters & PLCs
No cloud APIs · no external AI calls · no data egress
Connectors

Speaks database and machine

25+ connectors, including 10 industrial & streaming protocols. Every one performs a real protocol handshake on setup — with live read testing for industrial endpoints.

PostgreSQL MySQL SQL Server Oracle Snowflake BigQuery TimescaleDB InfluxDB MongoDB Modbus TCP/RTU OPC-UA BACnet/IP MQTT SCADA Apache Kafka OSIsoft PI + more
FAQ

Frequently asked questions

No. Every component — including the AI language model and vector search — runs as Docker containers on your own server or private cloud. There are no cloud API calls, no OpenAI keys and no per-token fees. Nerqon can run fully air-gapped, with no internet connection at runtime.

Nerqon runs on modest hardware — from a single server with 32 GB RAM. A GPU is optional and auto-detected: it speeds up AI answers but isn’t required.

Nerqon ships as Docker containers and deploys with one command — docker compose up -d. From bare server to running product typically takes under 15 minutes.

25+ connectors: relational databases (PostgreSQL, MySQL, SQL Server, Oracle…), cloud warehouses (Snowflake, BigQuery, Redshift, ClickHouse), time-series stores (TimescaleDB, InfluxDB, QuestDB, Prometheus), MongoDB — plus 10 industrial/streaming protocols including Modbus TCP/RTU, OPC-UA, BACnet/IP, MQTT, SCADA, Kafka, REST APIs, OSIsoft PI and AVEVA Historian. Full list →

You ask in plain language. The AI writes and runs read-only SQL against your databases, then answers with text plus the result table, an instant chart, a CSV download and the exact SQL it used. It knows your schemas and built-in energy standards (BEE/ASHRAE benchmarks), and it’s powered by a local 14-billion-parameter model — grounded in your database, not the internet.

Request a demo and we’ll contact you personally to arrange a guided tour on your screen — dashboards, the chatbot answering real questions, and forecasting. Demo access is provided individually after contact.

Ready to ask your data
where you’re wasting energy?

Book a personal demo — we’ll walk you through live dashboards, the AI chatbot and forecasting, tailored to your industry.