Changes released in July 2026.
"My Support Requests" in your user menu lists every request you have opened. Open one to read the team's replies, write back and attach a file. You can see when your message was read. New messages reach you in your notifications and by email, and replies appear while you have the thread open.
On "Parks Overview", each plant now shows how many of its inverters are producing. An orange line counts the inverters that have sent nothing for over an hour, a red line those that have stopped producing. Select a line to see which inverters they are. Each plant's type and peak power now sit under its name.
If a direct marketer curtailed only part of your plant, your lost energy was counted twice: against the marketer and the grid operator. It now counts once, from that part's share of expected production, and the grid operator's figure drops. Figures already issued keep their value until the period is rebuilt. Your plants receive this with their next agent update.
A plant's analysis now opens on a "Stations" tab. You see one row per data logger, whether it is still reporting, and how many of the inverters behind it are producing. Select a station to drill into its inverters, combiner boxes and strings. Switch on "Show sum of component" to see a whole level as one curve.
The assistant now does more than answer questions. After asking you first, it opens, updates, closes or reopens a ticket, comments on it, and links the components and events it concerns. It can also read your notifications and your activity, look things up in the documentation, and change your own settings.
Notices about your own account — sign-in, password, two-factor, sessions and API tokens — now reach you by email whatever your notification settings say. In your profile's "Notifications" tab they show as "Always delivered" and can no longer be switched off.
When an inverter stops producing, your on-site agent now raises a "Component production outage" event on your "Activity" page. It appears after three five-minute windows with no output, and only in daylight when snow, fog or curtailment does not explain it. It closes once production resumes. Your plants receive this with their next agent update.
If you pull raw data yourself, the old setting that combined plants across a portfolio is gone. You choose instead how several plants are put together: one series per measurement for all of them, or one series per plant, named after it. On the export page you make that choice with "Merge" or "Split".
Action required: API users: raw data export calls no longer accept aggregation, start_at_zero or delta — these parameters are ignored. Send value=plain, value=start_at_zero or value=use_delta instead of the two booleans, and multi_park_agg=merge or multi_park_agg=split to choose how several plants are combined. For metrics exports, move to park, portfolio and year with quarter, month, week or day; timeframe, period, park_uids and portfolio_uids are still accepted.
Larger SUN2000 inverters on a Huawei SmartLogger reported only their first 12 or 18 strings, leaving the rest with no voltage and no current. Every wired string is now monitored, whatever the model, so your string figures cover the whole inverter. Your plants receive this with their next agent update.
An on-site analysis now tells you when your inverters take in more DC power than their listed strings account for. You see how much is unaccounted for, and roughly how many strings that is — either missing from the configuration, or not reported by the data logger. Your plants receive this with their next agent update.
If you call the raw data export yourself, you now name the raw metrics you want and Mirox does the rest. Query text you send along is no longer run, and anything that is not one of the raw metrics is refused. Exports you build on the export page are unaffected.
Action required: API users: change your raw data export calls to send metric_ids — one to twenty raw metric ids, which you can read from the raw metrics list the API offers. A list of queries is still accepted, but only each entry's id counts and it must be a raw metric id; anything else is refused, and the whole export fails. Query text you send is no longer executed.
On a plant's production page, the lost energy shown for a single day was too high and did not match the week and month bars. The day now uses the same measured value as those bars, so the figure falls to the one the longer periods already showed.
When someone invites you to their organization and you already belong to one, you can now accept from your profile. Mirox tells you it will remove you from your current organization first. If you are its last admin, that organization is deleted with its portfolios and plants instead, and you confirm by typing DELETE.
A new "Tickets" page lists the tickets of every plant you can reach. Open one to describe a problem, assign it to a colleague and discuss it in comments. Mention people with @, and link the components and events the ticket concerns. You can also start a ticket from a component's row.
The bell in the header now tells you what happened on your plants, your tickets and your organization. Select an entry to go straight there. In your profile's "Notifications" tab you choose which kinds of news reach you, from which priority on, and whether they also reach you outside the app.
A new "Activity" page gathers the events of every plant you can reach: what happened, who or what caused it, and when. A grid of the last 30 days shows which days were busy. You can filter the feed by day, priority or type, and open a ticket from an event.