The category is a solved product with a dozen good versions. This one is built on the opposite premise - not a feature you could bolt onto the others.
The category already works
There's a mature market of garden-planning apps, and most of them are good at what they do. They draw a bed to scale, snap plants onto a spacing grid, read your frost dates off a nearby weather station, and hand you a planting calendar with reminders. If that's what you want, several of them will serve you well, and some are cheaper and have bigger plant libraries than this one.
This planner isn't trying to win that contest on its own terms. It's built on a different premise, and the difference runs deeper than any single feature - you can't bolt it onto one of the others, because it's a decision about what a recommendation is even allowed to be.
The one difference: every suggestion has to show its work
In most planners, a recommendation is just an assertion. "Plant these together." "Start indoors six weeks before your last frost." "These are good companions." The advice might be perfectly sound, but you're asked to trust it because it's on the screen, and there's no way to ask why, or how sure, or who says so.
Here, every rule the engine fires has to name the reason it works in one sentence, carry an evidence grade, and cite its sources - or it gets downgraded until it can, or dropped. When this planner tells you not to do something, the reason is right there, gradable and checkable:
PromisingDo not establish Verticillium-susceptible perennials (raspberries, strawberries, stone fruit) in ground that recently hosted Verticillium dahliae — either nightshades, or a logged planting of a Verticillium host (strawberry, raspberry or other bramble, or stone fruit).
costly if ignored → warn
Mechanism: Verticillium dahliae persists in soil as microsclerotia and infects a wide host range spanning nightshades and rose family. There is no in-ground cure once a susceptible perennial is planted, so it can wilt and die from inoculum an earlier host left behind — and because the reservoir family may differ from the victim's, the same-family rotation intervals miss it.
Evidence: confirmed against 3 sources in the research-based literature.
the 3 sources
Penn State Extension, 'Bramble Disease - Verticillium Wilt' (Kathy Demchak), extension.psu.edu/bramble-disease-verticillium-wilt (read 2026-07-12): 'Verticillium wilt is caused by two common soilborne fungi (Verticillium dahliae and Verticillium albo-atrum) that have a wide host range'; 'To minimize disease, choose a planting site with no known history of this problem. Avoid land recently planted with tomatoes, potatoes, eggplants, peppers, strawberries, raspberries, or stone fruits'; 'There are no effective fungicides for management once the plants are in the ground'; 'The number of years required to eliminate verticillium, especially the resting spores from the soil, is unknown.'
UMass Extension, New England Small Fruit Management Guide, 'Brambles - Diseases', ag.umass.edu/fruit/ne-small-fruit-management-guide/brambles/diseases (read 2026-07-12): 'Verticillium attacks a wide range of plants, including potatoes, tomatoes, peppers, eggplant and strawberries. It overwinters in the soil and on plant debris. Do not plant [raspberries following these crops].'; 'Non-host crops such as corn or wheat can help eliminate the fungus if grown for at least 2 years before planting raspberries.'
Host range / persistence corroborated by the source already cited on another rule here: NC State Extension, 'Verticillium Wilt of Tomato and Eggplant' (read 2026-07-11): 'The pathogen can survive in the soil without a host for up to 10 years in the form of structures called microsclerotia'; 'affects more than 300 host families.'
remedy
Site the perennial on ground with no recent nightshades, or choose one that is not Verticillium-susceptible. Extension advises growing a non-host — corn or a small grain — for at least two years first, without claiming that clears the soil. Once the perennial is planted there is no in-ground cure.
Read the full rule → · open in the planner ↗
That's not a fact box bolted onto the advice. It is the advice. The whole corpus - 92 rules today - is built to that standard, and a claim that can't meet it doesn't get to become a recommendation.
What that premise forces everywhere else
Once "show your work" is the rule, a lot of the category's standard furniture becomes impossible to build honestly, and this planner goes without it on purpose:
- No companion-planting chart. The grid of good and bad neighbours can't give a reason for each cell, so almost none of it survives the test. The long version is its own page; the short version is that 35 of the individual claims are kept as graded records instead, each with our answer.
- No invented numbers. When a recommendation would need a soil pH or a fertiliser figure the corpus doesn't have, the planner says "get a soil test" rather than guess. A number you read here is a number something actually measured.
- Odds, not a date. Your frost date is really a probability, so this planner shows you the band - half the years frosted by this date, one year in ten as late as that one - instead of one confident day. Why a hardiness zone can't give you a frost date is its own explainer.
- Rotation that survives a redesign. History sticks to the ground, not to a bed you drew, so rearranging the yard next spring doesn't wipe out what grew where. Most tools root history at the bed, which is why they forget.
- No AI making things up. There's no chatbot inventing planting advice here. An answer with no reason and no source is exactly the thing this corpus exists to refuse; a fluent guess is still a guess.
Your garden is yours, including when you leave
One more difference, with nothing to do with the corpus and everything to do with trust: your data never becomes a hostage. Your whole season log exports as a plain text file you can read, keep, and paste back in - no account needed to plan, and nothing you entered is locked behind a subscription that stops working the day you cancel. The planner is free for planning; if a paid tier ever shows up it'll be for convenience, never for access to your own records.
What it costs you
Here's the honest part, because the premise has a price and pretending otherwise would break the whole point. This planner has a smaller plant library than the big commercial tools - 83 species, on purpose, because a graded record for each one is slower to build than a shallow entry, and the corpus would rather be short and right than long and padded. It has no companion chart, which some people genuinely want. It won't hand you a confident single date when the truth is a range. And it grows more slowly, because a human has to read a source before a claim can be called verified.
If you want the biggest library and the most features for the least money, one of the others is probably your tool. If you want a planner that can always tell you why it said what it said, and show you exactly how sure it is - that's the one thing this one is built to do, and they aren't.