A standalone Streamlit app for comparing LLM responses across different user persona configurations. No FastAPI backend or MongoDB required.
pip install -r requirements.txtOPENROUTER_API_KEY=sk-or-v1-your-key streamlit run app.py| Variable | Default | Description |
|---|---|---|
OPENROUTER_API_KEY |
(required) | Your OpenRouter API key |
MODEL_NAME |
google/gemini-2.5-flash |
Model to use via OpenRouter |
- No FastAPI dependency - the app calls the LLM directly via
pydantic-ai+ OpenRouter. - No MongoDB - chat history is kept in Streamlit session state (in-memory, per browser tab).
- Persona prompt built locally - dimension descriptions are embedded in the app instead of being looked up from a database.