Wireless sensors and Yggio, the open operational layer for cities, property owners and transport operators. Your data, your operations, your sovereignty.

A person walking up a stairway into a glowing city of data
Home›AI›Yggio AI
Yggio AI

What if your city had its own AI?

An AI that knows what the large language models know, and also what is happening across your operations right now: every building, pipe, vehicle, machine and sensor, in real time.

How data-driven could your operations become? How much safer, more secure, with how much less CO₂?

What if… the fire engine was routed around today’s road works before it left the station?
See what Yggio AI can do ↓ Talk to us
The operational layer

Yggio, the operational layer, empowered by AI

Between the physical city and the people and services that run it sits the operational layer: every sensor, network and system connected once, translated into one data model and governed by one set of rights.

Put AI in that layer and it sees the whole city, not one silo at a time. It can read every data source it is allowed to, act through the same rules as everything else, and be switched off with one click.

Sovereignty and data quality

Your AI. Your data. Your rules.

AI is only as good as the data it reads and only as safe as the access it is given. With Yggio you own both, and you can prove the data is right before anyone acts on it.

Your environment

  • Yggio runs where you want it: as a service (by default with our partner Cleura for the EU market), on your own servers or at the edge
  • Local language models for maximum sovereignty, so no data has to leave your environment
  • All connections outbound; no inbound port is opened

Your rights

  • The access rights manager decides which data every person, service and AI agent can reach
  • Admin, Read, Write or Peek per resource, inherited down the organisation tree
  • Every right can be withdrawn at once

Your data quality

  • 500 translators turn every device into one NGSI-LD data model, with the same names and the same units
  • Metadata on the data, so a person or an AI knows what a value means and where it came from
  • AI-based anomaly detection flags a sensor that starts reporting differently from its neighbours or its own history
500translators into one data model
1NGSI-LD data model for every vendor
4access levels per resource, for people and agents alike
0critical, high or medium findings in an independent penetration test, January 2026
AI capabilities

Five ways to put AI to work in Yggio

From AI models trained on your own data to an assistant that knows Yggio, and next one that answers from your live data. All inside the same platform, the same rights and the same rules.

AAI modelsBMCP serverCAI agentsDHelp chatbotEComing next · Chat with your data
Yggio dashboard for a building run by an AI model: energy, money and CO₂ saved, occupancy, set points and outdoor conditions
A
AI models

AI models trained on your own data

Yggio trains an AI model on a sensor’s own history and scores every new reading against it, without the data leaving the platform. Training and inference run as ordinary Yggio services next to the data, so there is no export step and no separate machine-learning stack to run. AI Model Deployment Platform in Sensative Docs.

  • Anomaly detection per sensor, on any numeric measurement: a leak, a blocked filter, failing equipment, a door opening at an hour it normally does not
  • Yggio trains several candidate models, keeps the best and records how they scored, so a model’s provenance can be shown later
  • One choice is left to you: how often an anomaly should be expected, which in effect sets how many alarms you will act on
  • Results arrive as ordinary readings, for dashboards, alarms, the Rule Engine, export and access rights
In productionOne AI per building plans heating and ventilation, and a supervisor keeps every value within your limits.
AI for Buildings
The Yggio access rights manager: members of an organisation unit with Admin, Read, Write and Peek rights per resource
B
MCP server

An MCP server that lets you decide what AI can reach

The MCP server is how AI reaches Yggio. Today it gives the AI Assistant the documentation; next it opens your device data, and lets an assistant carry out work in Yggio through the API. What any assistant or agent can see and do is decided by you, in Yggio’s access rights manager, down to a single device. AI Connector in Sensative Docs.

  • An agent is a member of the organisation tree, like any user or service
  • Admin, Read, Write or Peek per resource, and nothing beyond what you grant
  • Withdraw access the same way as for a person, effective immediately
Yggio navigation for a rescue vehicle: a route around two blocked areas marked on the map, with car or truck and one-way override options
C
AI agents

AI agents that deliver services

Agents in Yggio do work that used to need a person: they read, interpret and act on information as it arrives.

  • An Agent connector runs a language model with a fixed system prompt and a chosen response format (text, JSON or a JSON Schema), so its answer can feed the next step automatically
  • OpenAI or Claude (Anthropic), with a token cap per response and per day
Rescue services, GotlandAn agent reads road-works plans and routes the fire service around closed roads. This is now part of standard Yggio.
The Gotland case
The Yggio AI Assistant answering questions about Yggio in the control panel sidebar
D
Help chatbot

An AI Assistant that knows Yggio inside out

A chat in the control panel, built on Yggio’s MCP server, that answers “how do I…” questions from the documentation about Yggio, our sensors and use cases. AI Assistant in Sensative Docs.

  • Answers how to connect a device, build a dashboard, set up a rule or a scheduled report, or find out why a device stopped reporting
  • Runs on your own Anthropic (Claude) or OpenAI account, with your own API key
  • Up to ten models per connector, with the temperature and a token cap per response set by you
  • A daily token limit per connector, 500,000 by default, that you set and can see
  • It answers from the documentation, not from your data; for that, see the next step
The Yggio Data Analyst Bot on a phone, answering questions about air-conditioning electricity use with a correlation and a plot by hour
E
Coming next · Chat with your data

Ask your data and your devices

The next step on the roadmap: the assistant also reads your device data and carries out work in Yggio through the API, within the rights you have granted. Ask a question in plain language and get an analysis, a chart or a table back.

  • An instant data analyst on Yggio’s time series, combined with other sources when you want
  • On a computer, a phone or by voice
  • Supports local language models, for maximum sovereignty

What would your city ask its own AI?

Tell us what you want AI to do with your city, building or fleet data. We will show you how far Yggio gets you today and start with a pilot where it counts.

Talk to us AI Research