Prove Stella onyour plant, your history,before you commit to a year.
A paid 60-day pilot for SMT and EMS plants of fifteen lines or fewer. She reads about sixty days of your history first, walks your team through seven mornings before go-live, then runs every day. At day sixty, you decide whether she stays.
One plant. Every line she can read. Sixty days.
The pilot is priced at the monthly plant-level rate for the tier you choose. No setup, onboarding, per-user, or per-line charges.
One plant, one decision maker. Priced at the plant level, not per line.
Strongest between three and eight SMT lines. Holds well to about fifteen.
Long enough for her to earn a place in your morning, or not.
Start with your own mornings, not a transformation program.
You don't need a software initiative to find out if this works. The pilot answers one question: does your team start the day with her, and does what she says hold up on your floor? Your data stays yours. She reads the exports you choose to provide and connects to nothing.
A bounded pilot built around real daily use.
Everything below is included in the pilot fee. Onboarding is run remotely by Stellarus.
Morning Brief
Every morning before first shift: what matters most, why, and what to do first, with the evidence behind each finding.
Whole-plant reading
She reads every export you give her: downtime, changeover, work orders, quality, test, materials. She tells you what she watched and what she didn't.
Evidence on every finding
Every claim she makes traces to rows in your own record. Nothing is said that the data can't show.
Mapping from your data
We map your exports as they already are. Questions that need plant knowledge come to you; everything else is resolved from the record.
Seven walk-through mornings
Before go-live, your team reads her briefs on mornings they remember and judges them against what actually happened. Ten minutes each, remote.
Day-60 decision
You decide whether she converts to an annual plan. Either way, you get an export of your data and everything your team taught her.
What she needs from you, and what she doesn't.
Nothing to install, nothing to connect, nothing to replace.
What she reads
The daily exports your MES, quality, test, and downtime systems already produce. CSV is enough.
What she doesn't need
No MES replacement, no database access, no real-time integration, no hardware. She connects to nothing and changes nothing on your floor.
How much history
About sixty days of exports to set her baselines. After that, one export a day.
Whose data it is
Yours. Stored per plant, isolated from every other plant, never resold, exported back to you on request.
Where the model sits
The language model never sees your raw data. It reads only what the engine found.
Bad days
Late files, duplicates, partial days, and backfills are handled. A day with no export is a day she tells you she didn't see.
How the 60-day pilot works.
You provide about sixty days of historical exports in the formats your systems already produce. Nothing to buy or build.
We map your data through onboarding and resolve what we can from the data itself. Only questions that need plant knowledge come to you.
Seven short remote sessions, one per day. Your team reads what she says about mornings they remember and tells us where she's right and where she isn't.
Your plant provides each day's export. Her brief is ready before first shift. At day sixty, you decide.
By the end of the pilot, you should know four things.
Prove she's right on your floor before you commit to a year.
See if your plant is a fit.
Fifteen SMT lines or fewer, one plant with one decision maker, and an owner who believes technology matters. If that's you, let's talk about a pilot.