PlantOntology

The world's first open-source plant knowledge graph — built by landscape professionals, for everyone.

plantontology-production.up.railway.app

PlantOntology is an open-source knowledge graph that models plants, ecosystems, and landscape design relationships — making 19 years of professional landscape expertise available to everyone through AI.

🌍 Why PlantOntology?

Existing plant databases (iNaturalist, GBIF, Plants.com) store facts — scientific names, photos, basic traits.

PlantOntology stores relationships and intelligence:

수종 ──[동반식재]──▶ 수종          # Companion planting
수종 ──[기피관계]──▶ 수종          # Allelopathy / conflict
수종 ──[기후적합]──▶ 기후존        # Climate zone match
수종 ──[유지관리]──▶ 난이도        # Maintenance complexity
수종 ──[생태역할]──▶ 조류/곤충     # Ecological function
수종 ──[병해충]───▶ 방제법         # Pest/disease management
포장 ──[심미조합]──▶ 수종군        # Aesthetic combinations
토양 ──[적합수종]──▶ 수종          # Soil-species match

This powers questions like:

• "느티나무 옆에 뭘 심으면 좋을까?" → companion planting recommendations

• "서울 기후에서 자생종으로 공원 식재 계획 짜줘" → AI-generated planting plans

• "이 배치가 생태적으로 건강한가?" → ecosystem health scoring

• "드라이가든용 내건성 수종 20종 추천" → climate-adaptive selection

Related MCP server: Neo4j MCP Server

🎯 Who is this for?

🗂️ Project Structure

PlantOntology/
├── data/
│   ├── species/           ← Plant species JSON (Korean + global)
│   ├── relationships/     ← Companion, conflict, ecological edges
│   ├── climate_zones/     ← Korean climate zones (쾨펜 분류)
│   └── regulations/       ← 도시공원법, 건축법 조경 기준
├── ontology/
│   ├── schema.cypher      ← Neo4j node/relationship schema
│   └── constraints.cypher ← Unique constraints & indexes
├── api/
│   ├── main.py            ← FastAPI app
│   └── routers/           ← /species, /recommend, /plan, /ecosystem
├── sdk/
│   └── plantontology/     ← Python SDK (pip install plantontology)
├── mcp/
│   └── server.py          ← MCP server for Claude Code integration
└── scripts/
    ├── seed_neo4j.py      ← Load initial dataset into Neo4j
    └── import_gbif.py     ← Import from GBIF open data

🚀 Quick Start

git clone https://github.com/alexai-mcp/PlantOntology
cd PlantOntology
pip install -e ".[dev]"

# Start Neo4j (Docker)
docker run -p 7474:7474 -p 7687:7687 neo4j:latest

# Seed initial dataset
python scripts/seed_neo4j.py

# Start API
uvicorn api.main:app --reload

Use as MCP Server (Claude Code)

{
  "mcpServers": {
    "plantontology": {
      "command": "python",
      "args": ["-m", "mcp.server"],
      "cwd": "/path/to/PlantOntology"
    }
  }
}

📚 Documentation

• Getting Started — Setup and first queries

• Contributing Guide — How to contribute

• Changelog — What's new in each release

• OpenCrab Grammar — 9-space semantic architecture

• API Reference — REST API endpoints (coming soon)

• Neo4j Setup — Full database configuration

🌱 Initial Dataset (Korean Native + Ornamental)

Phase 1: 500 species with full relationship data

• 한국 자생 수종 200종 (Korean native trees/shrubs)

• 조경 활용 교목/관목 200종 (Common landscape species)

• 지피식물/초화류 100종 (Ground covers & perennials)

Each species includes:

• 학명 / 국명 / 영명

• 생육 특성 (수고, 수폭, 생장속도)

• 기후 적합성 (한국 기후존 1–7)

• 토양 적합성

• 유지관리 난이도

• 동반식재 / 기피 관계

• 생태 기능 (탄소흡수, 조류유인, 밀원)

• 조경 활용처 (가로수, 공원, 정원, 옥상녹화)

🗺️ Roadmap

🤝 Contributing

PlantOntology thrives on contributions from:

• Landscape architects — domain knowledge, species data

• Botanists — ecological relationships, taxonomy

• AI engineers — Graph RAG, recommendation algorithms

• Translators — Korean ↔ English ↔ Japanese species data

• Gardeners — real-world companion planting observations

See CONTRIBUTING.md for guidelines.

📄 License

MIT — free to use, modify, and distribute.

Built with 19 years of landscape expertise + AI by AlexLee Powered by Neo4j · FastAPI · Claude Code MCP