Antti Haaraniemi

DataPortia - On-premises industrial data acquisition with local AI

by
DataPortia is on-premises industrial data acquisition software with built-in AI analytics — no cloud required. Connects to any OPC UA automation system, collects 2000+ measurements per second into TimescaleDB, and analyzes them locally using Ollama LLM. Five AI analysis types: anomaly detection, forecasting, cost optimization, reports, and natural language queries. Real-time dashboards, interactive trends, automated PDF/CSV reports, and OPC UA alarm management. Your industrial data is safe.

Add a comment

Replies

Best
Antti Haaraniemi
Hey Product Hunt! 👋 I'm Antti, the maker of DataPortia. **The problem I wanted to solve:** I've worked with industrial automation systems where plants generate thousands of measurements every second — temperatures, pressures, flow rates, energy consumption. But getting that data out of the SCADA/PLC systems and into something useful? That usually means either expensive proprietary historian software (Simatic Process Historian, Wonderware, AVEVA) or cobbling together Excel exports and manual reports. And when it comes to AI analytics on industrial data? Every solution wants you to send your sensitive process data to the cloud. For many industrial facilities, that's simply not an option — cybersecurity policies, air-gapped networks, data sovereignty requirements. **What DataPortia does differently:** - Connects to any OPC UA automation system and collects 2,000+ measurement points per second - Stores everything in TimescaleDB (handles 172 million rows/day, compresses efficiently) - Runs AI analysis *locally* via Ollama — anomaly detection, forecasting, cost optimization — your data never leaves your network - Modern web UI with dashboards, interactive trends, automated reports, and alarm management - 4 languages built in: Finnish, English, Swedish, German **How it evolved:** DataPortia started as a simple data logger. Over time, customers needed trends, then reports, then dashboards. The biggest "aha moment" was when I realized that the same Ollama models I was experimenting with could analyze industrial process data directly — no cloud API calls, no subscriptions, no data leaving the plant. The whole stack runs on open-source technology (PostgreSQL, TimescaleDB, Ollama) instead of proprietary databases, which keeps the total cost of ownership dramatically lower. **I'd love your feedback on:** - The AI analysis approach (local LLM vs. cloud) — does this resonate? - What features would you expect from industrial data software? - Any UX thoughts from the screenshots? There's a free 30-day trial with full functionality if you want to try it out. Happy to answer any questions! 🚀