EdikteFinder Analyzer

A desktop application for discovering, downloading, and AI-analyzing Austrian judicial real-estate auctions (Gerichtsversteigerungen) from edikte.justiz.gv.at.

⚠️ Legal Disclaimer: This tool is for research purposes only. AI-generated analyses are not financial or investment advice. See DISCLAIMER.md for the full disclaimer in English and German.

---

Features

• Automated scraping – searches the Austrian Edikt portal using simple, Aktenzeichen, or extended search modes; enriches every result with full detail-page metadata
• PDF download – fetches the expert appraisal (Langgutachten) directly from the court portal
• Multi-provider AI analysis – sends the PDF text + Edikt metadata to your chosen AI and returns a structured investment assessment:
  • Investment score 1–10 (decimal)
  • Risk class & return potential
  • Up to 6 concrete opportunities and 6 risks
  • Market position, renovation cost estimate, detailed summary
• Five AI providers – OpenAI, Anthropic Claude, Google Gemini, xAI Grok, or a local Ollama model (free)
• 100 % local / serverless – no database, no server, no cloud storage; all data lives in data/jsons/ as plain JSON files
• Dark-mode PyQt6 UI – search panel, results table, detail panel, investment overview tab with KPIs

---

Screenshots

(Add screenshots here after first run)

---

Requirements

Dependency	Version
Python	≥ 3.11
PyQt6	≥ 6.7
Playwright (Chromium)	≥ 1.44
pdfplumber	≥ 0.11
httpx	≥ 0.27

AI provider libraries are optional – install only what you use:

Provider	Package
OpenAI / Grok	openai
Anthropic	anthropic
Google Gemini	google-genai
Ollama	(no extra package – uses httpx)

---

Installation

# 1. Clone
git clone https://github.com/JakOb-dotcom/EdikteFinder-Analyzer.git
cd EdikteFinder-Analyzer

# 2. Virtual environment (recommended)
python -m venv .venv
.venv\Scripts\activate        # Windows
# source .venv/bin/activate   # macOS / Linux

# 3. Install Python dependencies
pip install -r requirements.txt

# 4. Install Playwright browser
playwright install chromium

# 5. Create your settings file from the template
copy data\jsons\settings.example.json data\jsons\settings.json
#   → then edit settings.json with your API keys (see Configuration below)

---

Configuration

All settings are stored in data/jsons/settings.json (excluded from Git). You can also edit them through the app's ⚙ Einstellungen dialog.

{
  "ai_provider": "ollama",
  "openai_api_key": "sk-...",
  "openai_model": "gpt-4o-mini",
  "anthropic_api_key": "sk-ant-...",
  "anthropic_model": "claude-haiku-20240307",
  "gemini_api_key": "AIza...",
  "gemini_model": "gemini-2.0-flash",
  "grok_api_key": "xai-...",
  "grok_model": "grok-3-fast-beta",
  "ollama_base_url": "http://localhost:11434",
  "ollama_model": "llama3.2",
  "max_context_chars": 40000
}

Tip for free local analysis: Set ai_provider to "ollama" and run Ollama locally – no API keys, no costs. Recommended model: qwen2.5:32b or llama3.2 for German-language documents.

---

Usage

python main.py

1. Search – select a search mode (simple / Aktenzeichen / extended), fill in the filters, click Suchen. The app fetches all result detail pages automatically.
2. Download – select one or more entries and click ⬇ Gutachten laden to download the expert appraisal PDF.
3. Analyze – click ✦ KI-Analyse to send the PDF text to your configured AI provider. Results appear in the detail panel and the Investitions-Übersicht tab.

---

Project Structure

EdikteFinder-Analyzer/
├── main.py            # PyQt6 UI – all windows, panels, worker threads
├── scraper.py         # Playwright scraper for edikte.justiz.gv.at
├── ai_analyzer.py     # AI backends (OpenAI, Anthropic, Gemini, Grok, Ollama) + PDF extraction
├── storage.py         # JSON-based persistence (edikte.json, analyses.json)
├── config.py          # Central config, loads/saves settings.json
├── requirements.txt   # Python dependencies
├── data/
│   ├── downloads/     # Downloaded PDFs (git-ignored)
│   └── jsons/
│       ├── settings.example.json   # Template – copy to settings.json
│       ├── settings.json           # Your settings with API keys (git-ignored)
│       ├── edikte.json             # Scraped Edikt data (git-ignored)
│       └── analyses.json           # AI analyses (git-ignored)
├── LICENSE
├── DISCLAIMER.md
└── CONTRIBUTING.md

---

AI Provider Comparison

Provider	Cost	Privacy	German quality	Best model
OpenAI	Pay-per-token	Cloud	⭐⭐⭐⭐⭐	GPT-5
Anthropic	Pay-per-token	Cloud	⭐⭐⭐⭐⭐	Claude Sonnet 4.5
Google Gemini	Pay-per-token (free tier available)	Cloud	⭐⭐⭐⭐	Gemini3.0 Flash
xAI Grok	Pay-per-token	Cloud	⭐⭐⭐	Grok3
Ollama (local)	Free	100% local	⭐⭐	qwen2.5:32b
You can also try other models and easily change the model, just type in the model name and put in your API key, currently supported are all models from OpenAI, Anthropic, Google Gemini, xAI Grok and Ollama.

---

Legal & Privacy

• This software accesses publicly available data from edikte.justiz.gv.at. No authentication is bypassed.
• No personal data is collected by the application. All scraped data stays on your local machine.
• PDF files downloaded from the portal are court documents in the public domain.
• AI analyses are generated locally or via third-party APIs at your own cost and responsibility.
• See DISCLAIMER.md for the full legal disclaimer (English + German).

---

Contributing

See CONTRIBUTING.md. Bug reports and PRs are welcome!

---

License

MIT License – © 2026 JakOb-dotcom & EdikteFinder-Analyzer Contributors