mysmart.garden
"A gardener for people who kill plants."
I am a novice gardener - the kind who buys four pepper plants with real optimism and then stands in the yard a week later wondering if the wilted one is thirsty, drowning, sunburned, or just dramatic. So this spring I built mysmart.garden: a total garden-management app for my whole property, from the raised beds and tomato grow bags to the herb boxes and the strawberry jar that betrayed me last year.
The bet behind it: the right interface for a garden isn't a form, it's a conversation with something that already knows the answer.
Conversation First
The primary interface is chat. I describe my garden in plain English - "I put four mini bell peppers and four rainbow peppers in the long raised bed, and the end ones got frost-bitten" - and an AI agent turns that into structured data: plants, beds, positions, even a note about the freeze damage. Forms still exist, but only as an escape hatch when the AI gets something wrong. I describe; it files.
The agent has a toolbox of functions it can call - create a bed, move a plant, log a watering, analyze a photo, plan a season - so the conversation isn't just chat, it's the steering wheel for the entire database. Sprout, a seven-expression mascot, is the agent's face: it waves on the way in, looks concerned at a frost warning, and celebrates a logged harvest.
What It Does
- Morning briefing. A daily email that already knows the weather and fires per-person at each household member's chosen hour.
- Weather watching. Pulls the National Weather Service forecast, cross-checks it against Open-Meteo, and flags uncertainty instead of lying confidently. Overnight rain auto-logs the watering.
- Frost and heat alerts. Warns when peppers are about to get bitten - without spamming the same warning twice.
- Photo diagnosis. Snap a picture of a sad leaf in the chat and the agent reads it, then files it against that plant for a visual history.
- Adaptive watering. Weights days-since-watered by how hot it's been, so a stretch of 90° afternoons moves the needle faster than a mild week.
- Households and helpers. Data belongs to a household, not a login. When we travel, my father-in-law gets a one-button "Done" email to confirm he watered - no app, no account.
The Stack
- Next.js 16 (App Router, Turbopack) + React 19 + TypeScript
- Supabase - Postgres, magic-link auth, row-level security, household-scoped data
- Vercel AI Gateway via the Vercel AI SDK - Haiku 4.5 for fast work, Sonnet 4.6 for heavier reasoning
- Vercel Pro - hosting plus five scheduled cron jobs for briefings, rain auto-logging, alerts, and helper reminders
- smtp.com + Nodemailer - the email that makes the whole thing feel alive
- Tailwind v4 + Radix - a mobile-first UI on a tiny six-color palette and exactly two fonts
Where It's At
This is very much an active build. Onboarding, the weather pipeline, daily briefings, helpers, care tasks, photo analysis, and weather alerts are all shipped and running in production against my real garden this season. Illustrated scene generation works but is still rough, and the roadmap is long: a rooftop weather station instead of a 2km grid cell, richer onboarding, and eventually the flowers and bushes that turn this from a veggie tracker into a whole-property garden brain.
mysmart.garden is personal and single-user for now - it's gated behind magic-link auth with no public sign-up - so the link is a look at what it is rather than a place to make an account.
For the longer story behind it, read mysmart.garden: An AI Gardener for People Who Kill Plants.