AppArt Agent¶
AI-powered apartment purchasing decision platform for France
AppArt Agent helps buyers make informed real estate decisions by combining French property transaction data (DVF) with AI-powered document analysis and visualization tools.
Features at a Glance¶
flowchart LR
subgraph Input["Your Input"]
Address["Property Address"]
Docs["Documents<br/>(PV AG, Diagnostics,<br/>Tax, Charges, Other)"]
Photos["Apartment Photos"]
end
subgraph AI["AI Analysis"]
DVF["Price Analysis<br/>5.4M+ transactions"]
DocAI["Document Analysis<br/>Native PDF + Thinking"]
PhotoAI["Photo Redesign<br/>Style visualization"]
end
subgraph Output["Decision Support"]
Price["Market Valuation"]
Risks["Risk Assessment<br/>+ Confidence Score"]
Costs["Cost Breakdown<br/>+ Tantiemes"]
Visual["Renovation Preview<br/>+ Promoted Redesigns"]
end
Address --> DVF --> Price
Docs --> DocAI --> Risks
Docs --> DocAI --> Costs
Photos --> PhotoAI --> Visual
Price Analysis¶
Access 5.4M+ French property transactions from DVF data to understand market prices, trends, and get personalized recommendations.
Document Analysis¶
Upload PV d'AG, diagnostics, tax documents, charges, and other property documents (rules, contracts, insurance). AI automatically classifies, analyzes with native PDF processing and reasoning, and produces a cross-document synthesis with cost breakdowns, tantiemes calculation, risk factors, and buyer action items.
Photo Redesign¶
Visualize renovation potential with AI-powered style transformation of apartment photos. Promote your favorite redesign to feature it on the property overview.
Decision Dashboard¶
Property overview with AI synthesis summary, risk level badges, annual/one-time cost breakdowns, promoted redesign previews, and inline property editing.
Quick Start¶
# Clone and setup
git clone https://github.com/benjamin-karaoglan/appart-agent.git
cd appart-agent
cp .env.example .env
# Configure AI: set GEMINI_USE_VERTEXAI=true or add GOOGLE_CLOUD_API_KEY
# Start services (migrations run automatically)
docker-compose up -d
# Access
# Frontend: http://localhost:3000
# API Docs: http://localhost:8000/docs
Architecture Overview¶
flowchart TB
subgraph Client["Client Browser"]
Browser["Web Browser"]
end
subgraph Frontend["Frontend - Next.js 14"]
Dashboard["Dashboard<br/>+ Synthesis Preview"]
Documents["Documents<br/>+ Multi-phase Processing"]
Properties["Property Detail<br/>+ Inline Editing"]
Photos["Photo Studio<br/>+ Promoted Redesigns"]
end
subgraph Backend["Backend - FastAPI"]
API["REST API"]
AIServices["AI Services<br/>Native PDF + Thinking"]
DocProcessing["Document Processing<br/>+ Cross-doc Synthesis"]
end
subgraph DataLayer["Data Layer"]
PostgreSQL[("PostgreSQL<br/>5.4M+ DVF")]
Redis[("Redis Cache")]
Storage[("MinIO / GCS")]
end
subgraph External["External Services"]
Gemini["Google Gemini<br/>Vertex AI"]
end
Browser --> Frontend
Frontend --> Backend
API --> PostgreSQL
AIServices --> Redis
DocProcessing --> Storage
AIServices --> Gemini
Properties --> API
Deployment Options¶
flowchart LR
subgraph Local["Local Development"]
Docker["Docker Compose<br/>Free, all-in-one"]
end
subgraph Cloud["Production"]
GCP["GCP Cloud Run<br/>~$65-445/month"]
end
Local -->|"Ready for prod?"| Cloud
| Environment | Best For | Guide |
|---|---|---|
| Docker Compose | Development, testing | Docker Guide |
| GCP Cloud Run | Production, scaling | GCP Guide |
Technology Stack¶
| Layer | Technologies |
|---|---|
| Frontend | Next.js 14, React 18, TypeScript, Tailwind CSS, Better Auth, next-intl, pnpm |
| Backend | FastAPI, Python 3.10+, SQLAlchemy, UV |
| Auth | Better Auth (email/password + Google OAuth), HTTP-only session cookies |
| AI/ML | Google Gemini 2.0, Vertex AI, LangChain |
| Database | PostgreSQL 15, Redis 7 |
| Storage | MinIO (local), Google Cloud Storage (production) |
| Infrastructure | Docker, Terraform, GCP Cloud Run |
Documentation Sections¶
flowchart TB
Start["Getting Started"]
Arch["Architecture"]
Backend["Backend"]
Frontend["Frontend"]
Deploy["Deployment"]
Dev["Development"]
Start --> Arch
Arch --> Backend
Arch --> Frontend
Backend --> Deploy
Frontend --> Deploy
Deploy --> Dev
- Getting Started - Installation, prerequisites, and configuration
- Architecture - System design and data flow
- Backend - API reference, AI services, database models
- Frontend - Pages, components, and UI patterns
- Deployment - Docker and GCP deployment guides
- Development - Local setup, testing, and contributing
Support¶
- Issues: GitHub Issues
- License: Custom Non-Commercial License