Citizen science for plants that fell through the cracks
A lot of the plants we love have the same story: someone finds them, they hit the market, prices spike, Instagram goes wild… and almost no one collects real data about how they behave, how they reproduce, or what they need to survive back where they came from.
We preserve them on shelves and in cabinets, but not in a way that truly helps keep them in the wild.
PlantOS is trying to change that, without turning your hobby into homework.
Right now, for many rare and newly traded plants:
• We do not have proper timing data for growth, flowering or dormancy
• We do not have tested protocols for imports, acclimation or propagation
• We do not have reliable records of how different environments and feeds affect them
• We do not have much that scientists or conservation projects can actually use
Most of the knowledge sits in DMs, memory, half finished notes and old photos.
PlantOS is built to turn that everyday chaos into something structured enough that:
The core idea is simple: if you are already growing, watering, feeding, pollinating and importing plants, you are already running experiments.
PlantOS turns that into usable signal by:
in a way that can be compared over time
You do not need to write papers, understand statistics or run controlled trials. You just record what you are doing and what you see. The structure is what makes it science grade later.
We focus on things that are actually useful if you want to keep a species going long term.
All of this starts as your regular logs. Once aggregated and cleaned up, it becomes the kind of dataset that can actually inform protocols, ex situ collections and in situ work.
PlantOS is designed so that your data has a clear path:
plants, feeds, environment, imports, flowers, propagations — as detailed or as broad as you want
per plant, per plant type, per genus, per environment
candidate patterns reviewed before they become shared guidance
per species or hybrid, not per user — focused on timings, responses, protocols, not on you as a person
summaries, anonymised datasets, and candidate protocols can be shared with:
You are never expected to handle that part yourself. You just get the benefit of better, data backed guidance inside the app, and the knowledge that the same patterns can help outside your grow room.
Preservation does not mean tracking you. A few core principles:
Public profiles and public plants are opt in
Exact coordinates or personally sensitive details are not what we are after
We do not use your data to train OpenAI's general models
Research partners get behaviour of plants and lines, not who grew them
Your data should feel like a club where your plants are famous, not you.
This is not meant to be a one way street. As a user, you get:
✓ A real memory for each plant, not just old photos
✓ AI help that actually knows your plants, history and conditions
✓ Clearer, data backed care suggestions and feed recipes
✓ Transparency around imports, sales and provenance
✓ A feeling that your time and effort goes further than your own shelf
If you want to go deeper, there can also be:
PlantOS is not a magic bullet for conservation. It will not fix habitat loss or bad policy.
What it can do is:
Your collection was always more than just decor.
This is a way to treat it like that.
Training and running AI is not free for the planet. Data centers, GPUs and storage all have a real footprint. If we are going to use that power in the name of plant preservation, it has to be done with a bit of honesty and balance.
PlantOS is built with that in mind:
We aim to use efficient models and infrastructure, not the biggest thing we can rent. A lot of the useful work here is about clean structure and good data, not just throwing huge models at it.
Where we can choose, we prefer hosts and regions that run on a high share of renewable energy or have clear decarbonisation plans, instead of the absolute cheapest option.
As the platform grows, we plan to treat AI and server usage as a measurable cost, then offset it through actions that actually matter for plants. That can mean supporting habitat protection, restoration projects, ex situ collections or other work that improves survival odds in a concrete way, not just generic green stickers.
Any heavy AI feature should justify itself in terms of real value: clearer care, better protocols, stronger preservation signal. If it is mostly cosmetic or vanity, it is not worth the energy.
The goal is not to pretend there is zero impact.
The goal is to make sure that whatever energy PlantOS does use is paid back with practical gains for the plants and habitats we are claiming to care about.
You do not have to sign up for anything special to count. Using PlantOS as it is intended already contributes.