An intelligent Terminal User Interface (TUI) application that natively scrapes Bazos.cz classified listings and utilizes an autonomous LLM Agent (via OpenRouter/Google Gemini) to fetch, parse, and evaluate listings against any custom natural language objective.
- Agentic AI Search: Parses natural-language prompts like
"I want a used Lifepo4 battery for my 48v PV system"directly into optimal Bazos queries and categories. - Smart Evaluation: Downloads the listings and actively evaluates each one with Gemini, outputting a 1-10 relevance Score alongside a custom Description/Summary explaining the exact product state and why it matches your needs.
- Multi-Category & Pagination: Autonomously determines if a query needs to cross-search multiple categories at once. Click "Load More" to infinitely fetch down the index!
- Interactive TUI: Built on top of
Textual, featuring live background-processing data tables that natively handle mouse functionality and can instantly open Bazos links in your browser! - HTML Export Support: Generate standalone, interactive HTML and Plotly metric graphs outlining your fetched dataset with the click of a button.
- Clone the repository and initialize a virtual environment:
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt- Create a
.envfile in the root containing your API key:
OPENROUTER_API_KEY=sk-or-v1-xxxxxx- Boot the application:
./run.sh- Textual: Terminal Application GUI
- httpx: Asynchronous HTTP handling (Bazos JSON extraction)
- OpenAI / OpenRouter: Autonomous intent parsing and evaluation pipeline
- Pandas / Plotly / Jinja2: Rapid static data visualization rendering