PatternBase is a web app for designing permaculture gardens. Instead of organizing plants in rows, it models plant communities called guilds that have a fruit tree surrounded by nitrogen fixers, dynamic accumulators, ground covers, and pest confusers. You design the guild, observe outcomes over seasons, and your observations contribute to a shared evidence base anyone can learn from.
I built this because I couldn't find software that thought about gardens the way permaculture designers actually do: multi-year succession, functional plant roles, zone and sector analysis, whole-site systems. Most tools are either basic companion planting charts or enterprise ag platforms. I wanted something in between.
The technical pieces I found interesting: PostGIS powers climate matching, so you can find gardeners in your bioregion and see what's actually working in similar growing conditions. Plant observations are geolocated and aggregated so patterns emerge across climates without any manual curation. The guild designer is built on p5.js with procedurally generated visuals and no static assets, everything is drawn from data. AI photo analysis (Claude API) lets you log an observation from a phone photo and get a structured record back.
Stack: Next.js 15 App Router, Supabase (PostgreSQL + PostGIS + Auth), TypeScript, Tailwind, Serwist for PWA. Built entirely solo using Claude Code across 90 sessions over about six months.
Honest limitations: the plant database has around 4,500 species whereas competitors have 20K+. No native mobile app. I have zero brand recognition in the permaculture space. The climate matching is only as good as the observation density, which right now is thin.
Pricing: free tier is genuinely usable (3 gardens, full guild designer). $5/mo or $149 lifetime. 25% of revenue goes to ecological restoration. Bitcoin payments accepted with a 15% discount.
I'm not looking for signups from this post. I'd genuinely like feedback on: whether the guild-first mental model makes sense to someone unfamiliar with permaculture, what the onboarding breaks on, and whether the climate matching feature lands clearly.
james-pb•1h ago
I built this because I couldn't find software that thought about gardens the way permaculture designers actually do: multi-year succession, functional plant roles, zone and sector analysis, whole-site systems. Most tools are either basic companion planting charts or enterprise ag platforms. I wanted something in between.
The technical pieces I found interesting: PostGIS powers climate matching, so you can find gardeners in your bioregion and see what's actually working in similar growing conditions. Plant observations are geolocated and aggregated so patterns emerge across climates without any manual curation. The guild designer is built on p5.js with procedurally generated visuals and no static assets, everything is drawn from data. AI photo analysis (Claude API) lets you log an observation from a phone photo and get a structured record back.
Stack: Next.js 15 App Router, Supabase (PostgreSQL + PostGIS + Auth), TypeScript, Tailwind, Serwist for PWA. Built entirely solo using Claude Code across 90 sessions over about six months.
Honest limitations: the plant database has around 4,500 species whereas competitors have 20K+. No native mobile app. I have zero brand recognition in the permaculture space. The climate matching is only as good as the observation density, which right now is thin.
Pricing: free tier is genuinely usable (3 gardens, full guild designer). $5/mo or $149 lifetime. 25% of revenue goes to ecological restoration. Bitcoin payments accepted with a 15% discount.
I'm not looking for signups from this post. I'd genuinely like feedback on: whether the guild-first mental model makes sense to someone unfamiliar with permaculture, what the onboarding breaks on, and whether the climate matching feature lands clearly.
pattern-base.com