{"id":8581,"date":"2025-12-22T16:33:53","date_gmt":"2025-12-22T16:33:53","guid":{"rendered":"https:\/\/sensor-online.se\/?post_type=funktions&p=8581"},"modified":"2026-10-11T12:45:23","modified_gmt":"2026-10-11T12:45:23","slug":"ai-dataanalys-iot-saker-oppen-apis","status":"publish","type":"funktions","link":"https:\/\/sensor-online.se\/sv\/funktions\/ai-data-analytics-iot-secure-open-apis\/","title":{"rendered":"AI och dataanalys f\u00f6r IoT med s\u00e4kra \u00f6ppna API:er"},"content":{"rendered":"\n
\n
\n

AI and analytics<\/p>\n

AI and analytics for IoT sensor data through MCP<\/h1>\n

Let an AI model analyse your sensor data. Through an MCP server and the open API it reads structured data from Sensor-Online, finds patterns and deviations, and writes the report.<\/p>\n

Book a demo<\/a>Call +46 (0)500 6000 22<\/a><\/div>\n

500+ customers<\/b> \u00b7 100,000+ active sensors<\/b> \u00b7 Trusted by Akademiska Hus, Vattenfall, LKAB and HSB<\/p>\n<\/section>\n

How it works<\/h2>

From sensor data to insight.<\/p>

Sensor data<\/h3>

Readings are stored in Sensor-Online.<\/p><\/div>

API and MCP server<\/h3>

Structured JSON data is fetched through the open API.<\/p><\/div>

AI model<\/h3>

The model analyses the data, through n8n or directly.<\/p><\/div>

Insight<\/h3>

Comparisons, deviations and reports.<\/p><\/div><\/div><\/section>\n

What you can do<\/h2>

Questions that used to need an analyst.<\/p>

Compare week to week<\/h3>

For example compare energy use with weather conditions from one week to the next.<\/p><\/div>

Detect deviations<\/h3>

Let the system find readings that stand out from the normal pattern.<\/p><\/div>

Automatic reports<\/h3>

Have findings summarised and reported without manual work.<\/p><\/div>

Build on open data<\/h3>

The analysis uses structured data from the open API.<\/p><\/div><\/div><\/section>\n

Cloud AI or your own model?<\/h2>

Choose where the analysis runs.<\/p>

<\/th>Cloud AI model<\/th>Local LLM<\/th><\/tr><\/thead>
Connection<\/th>MCP connection to a cloud model, such as Claude, with n8n<\/td>MCP connection to an LLM you run yourself<\/td><\/tr>
Where analysis runs<\/th>In the AI provider\u2019s service<\/td>In your own environment<\/td><\/tr>
Best for<\/th>Getting started quickly<\/td>Organisations that already run a local LLM and want to keep data inside<\/td><\/tr><\/tbody><\/table><\/div><\/section>\n

A tool for people, not a replacement<\/h2>

Check results before you act.<\/p>

AI findings are a starting point. Review important deviations and recommendations with someone who knows the operation before acting on them.<\/p><\/section>\n

Related features<\/h2>

Secure IoT platform with REST API<\/h3>Read more \u2192<\/a><\/div>

Integrate IoT data with business systems<\/h3>Read more \u2192<\/a><\/div>

Node-RED and n8n automation<\/h3>Read more \u2192<\/a><\/div>

Alarm management<\/h3>Read more \u2192<\/a><\/div><\/div><\/section>\n

Frequently asked questions<\/h2>
How does AI connect to Sensor-Online?<\/summary>

Through an MCP server that fetches structured JSON data from the open API, and through n8n, so an AI model can analyse it.<\/p><\/details>

What can the AI do with my data?<\/summary>

Compare periods, such as energy use against weather from week to week, detect deviations and produce reports.<\/p><\/details>

Can I use my own AI model?<\/summary>

Yes. If you already run a local LLM, it can be used instead of a cloud model.<\/p><\/details>

Does my data have to leave my environment?<\/summary>

Not with a local LLM. With a cloud model, the data you choose to analyse is sent to that provider\u2019s service.<\/p><\/details>

Is this ready-made or custom?<\/summary>

It is a flexible capability built on the open API. Talk to us about the analyses you want.<\/p><\/details><\/div><\/section>\n

Want to put AI to work on your sensor data? Talk to an engineer.<\/h2>

Tell us what you want to find out, and we will show how it can be done.<\/p>

Book a demo<\/a>Call +46 (0)500 6000 22<\/a><\/div><\/div>\n

All trademarks belong to their respective owners.<\/p>\n