What energy data can be collected automatically, and how do you connect it to AI?

We get this question regularly, both from property owners and from companies building their own service on top of energy data: what data do you collect, and how do you get that energy data into an AI tool?

What’s stored in the data layer, and how you access it (the short answer)

Metry collects consumption and readings for electricity, district heating, district cooling, gas and water, as well as energy production from heat pumps and solar installations, data from charging stations and readings from climate and air quality sensors. All data is validated and quality assured before it is stored.

The data can be sent to an external system in three ways: through an open API, through ready-made integrations to energy management and sustainability systems, or through Metry’s MCP server, which makes the data available directly in AI tools that support the open MCP standard.

What energy data Metry collects

These are the media types we can help you collect today:

• Electricity
• District heating and cooling
• Gas
• Water (hot and cold)
• Energy production from heat pumps
• Net and gross production from solar installations
• EV chargers
• Climate sensors for temperature and humidity
• Air quality sensors for CO2 and TVOC

Metadata follows every meter: name, address, EAN and energy type. On top of the meters sits the structure you build yourselves, meaning how the meters relate to regions, properties and buildings. That structure is often what decides what you can use the data for – like asking an AI tool for data analysis covering a particular block or a particular type of building.

The data is collected from more than 1,000 connected data sources in northern Europe, and new ones are added continuously. Metry collects from utility providers’ customer portals and invoices, file transfers in CSV, Excel, XML and market-specific energy data formats, national energy data hubs, smart meters and gateways, SCADA and building control systems, cloud-based energy management systems, solar inverters and charging station suppliers.

Meters that cannot be connected can be read manually in our app Metry Readings.

Resolution and history depend on the data source

Metry delivers the resolution available in the data source, and we have built collection methods to handle different kinds of flows. Monthly values are included with most utilities, and daily values are available from many sources. Hourly values are collected through national energy data hubs and through sub meters. Several methods are usually used in parallel within the same portfolio, depending on what you need for your energy management or sustainability reporting.

History works the same way; if it is registered with the utility, we can fetch it, however far back it goes. That’s often an unexpected bonus during acquisitions; when a new portfolio is connected, you might get several years of data to compare future improvements with!

These are the average times from a value being published in the source to it being available in Metry:

• Monthly values: within a day
• Daily values: the following day
• Hourly values from utilities: all values for a day are available the following day
• Hourly values from other sources: within an hour, often faster

What quality assurance does before a value is stored

Every value we collect is checked before it reaches the database. The anomaly check assesses whether a value is reasonable, both compared with similar meters and against the meter’s own history. Every value is logged with its source and timestamp, so you can always trace back lineage and origin for a specific value.

On top of that, our algorithms monitor the collection rate across the portfolio and fill gaps in the data automatically where possible. New meters added at your utilities lands in an inbox in the Metry portal, rather than passing unnoticed.

We also keep track of all your credentials and help you give the right property managers access to the right parts of the collection.

How you reach the data: API, ready-made integrations or MCP

Open API: Included for all customers at no extra cost. There are three common patterns: live calls when your application needs the data, recurring syncs that fetch whatever has arrived since the last one, or file exports in a format we decide together, delivered over HTTP Post, (S)FTP, SMTP, S3 or Azure Blob Storage.

If your customers have their own Metry accounts, access happens with OAuth against their account. The API documentation is openly available at metry.docs.apiary.io. You are responsible for the implementation.

Ready-made integrations: For the most common energy management, sustainability and finance systems there are connections activated with a single click.

The MCP server: As an independent data provider, we think it goes without saying that you should be able to work with your energy data in an AI tool as well! So we built an MCP server that lets you ask questions of your energy data, in the AI tool you have chosen. …but what exactly is an MCP?

Connect your AI tool without building an integration

MCP stands for Model Context Protocol and is an open standard for how AI tools connect to data sources. Instead of someone exporting a file and pasting it into a chat, the AI tool fetches the data straight from Metry when it needs it.

You add the address https://mcp.metry.io among the connections in your AI tool, sign in with your usual Metry account and ask your question in plain language. In most tools, the connection process is one field and one button – no developer is needed, and nothing is installed locally.

Access is read-only. Data can be read, but neither created, changed nor deleted. Authentication uses OAuth against the Metry account that already exists, and the AI tool sees exactly the data that account already has access to. No new permissions are granted, no API keys and no service accounts. The server runs on AWS in the EU and stores neither credentials nor conversation data.

What the AI tool reaches is consumption and readings, the meters’ metadata and your property and building structure. In practice that means questions like “which meters have missing values for January?” or “compare electricity use per property in 2026 against 2025” are answered straight against your own data, in the tool you have already chosen.

Want to know more about Metry’s data collection, or talk through what you could do with our MCP server? Book a demo and we will be in touch!