Four questions about a production process. Is it stable. Is it capable of meeting the specification. Which inputs actually move the output. What could go wrong. Gradora answers all four from real measurement data, on a Mac at your desk or an iPad on the floor, and shows exactly what produced every number it puts on screen.
Every statistic comes from a tested function, never from a language model. Rounding happens once, at the end, and a rounded value is never fed back into another calculation.
One engine, two destinations. Keyboard-fast on a Mac next to your spreadsheet, glove-friendly on an iPad next to the machine. iPhone is not a target.
No account, no analytics, no server. Gradora has nowhere to send a measurement even if it wanted to.
Five workspaces, in the order you work. Data in, a chart or a study out, then a report that goes to a customer or an auditor.
| Workspace | What happens there |
|---|---|
| Data | Import by paste, drag and drop, or file picker, always through a column-mapping step. A spreadsheet-like grid with typed columns, plus a shop-floor entry mode for one measurement at a time. |
| Charts | Control charts: I-MR, X-bar R, X-bar S, p, np, c, u. Capability analysis: Cp, Cpk, Pp, Ppk. Histograms, Pareto charts and run charts. |
| Experiments | Design of experiments: a factor and level setup screen, a randomized run sheet for entering results, and an analysis view with main effects, interactions and an ANOVA table. |
| Process | A process map, and an FMEA table where each row carries severity, occurrence and detection ratings. Rows link to the control chart that actually monitors that failure mode. |
| Reports | Assemble a study into an export built for a management review, a customer or an auditor, with every assumption and provenance note still on it. |
A quality engineer on a Mac, a spreadsheet of measurements open in the next window. Paste or import the data, run the analysis, read the result, export a page for an email or a management review. Keyboard-heavy, and impatient about it.
The same person, or a technician, standing at a machine with an iPad, entering measurements as parts come off the line, or working through a designed experiment's run sheet. Possibly wearing gloves, possibly in poor light. Large touch targets and high contrast matter here in a way they do not at the desk.
One codebase, two destinations, so a chart reads the same wherever it was read from.
Quality software today is mostly Excel add-ins priced $299 to $379 that look like they were designed in 2005, or Minitab, where the average business contract runs about $8,640 a year. Gradora is a native Mac and iPad app, priced once rather than every year.
The interface stays quiet on purpose: no gradient cards, no celebration animation, no progress gamification. Every computed number sits next to what produced it and the assumptions behind it. A control chart's limits come from the process; a specification's limits come from the customer; Gradora never lets the two blur into each other, on screen or in an export.
Capability on a process that isn't stable still computes, but the app withholds a verdict rather than hand over a number that wouldn't survive being checked. The R chart is always read first, and if it signals, the X-bar verdict says so rather than guessing. Nothing is ever asserted as capable on the app's own authority: every verdict is measured against the requirement you enter for that dataset. On supported devices, an on-device model writes the plain-language reading beneath a chart. It never computes a number itself, and it stays visibly separate from anything the software actually calculated.
$299 once. Not $299 a year.
Free to download, with a free tier that's capacity-limited rather than time-limited: enough to see a real chart on your own data, not enough to run your whole program on it. One unlock removes the caps.
There is a second, entirely optional purchase at $99 a year for continued updates. It is never required to keep using what you already unlocked. No advertising in any version, and no subscription for the core product.
In development, and honestly so. There is no build to download, no purchase to make, and the screenshots that would normally sit on a page like this do not exist yet because the screens they would show are still being built.
Before it ships, five real datasets get run through Gradora and through the established tools alongside it, and the comparison is published rather than summarised. There will also be a round of testing with people who run these studies for a living.
If that is you, and you want a look before it ships, or you think a default described here is wrong, write in. One person is building this, working from real SPC, DOE and FMEA problems rather than a feature list.
hello@no.spam.invalid.handfulapps.com