soul.py v2.0 Demo — RAG + RLM Hybrid

Live demo for soul.py v2.0-rag-rlm branch.

Three retrieval modes:

• Auto — router classifies each query, dispatches to RAG or RLM
• RAG only — semantic vector search (focused queries)
• RLM only — recursive synthesis (exhaustive queries)

v0.1 demo: https://soul.themenonlab.com v1.0 demo: https://soul-v1.themenonlab.com v2.0 demo: this app

Railway env vars required

Variable	Value
ANTHROPIC_API_KEY	your Anthropic key
QDRANT_URL	Qdrant Cloud URL
QDRANT_API_KEY	Qdrant API key
AZURE_EMBEDDING_ENDPOINT	Azure OpenAI endpoint
AZURE_EMBEDDING_KEY	Azure OpenAI key