$ rag KNOWLEDGE BASE
● —
[ llm provider ]
Installed Ollama models — click to use
Endpoint for Ollama or OpenAI-compatible providers.
✓ An API key is already saved.
0 = deterministic
[ ollama models ]
Manage models inside the local Ollama container.
Browse models at ollama.com/library.
Installed models
[ embedding device ]
Controls which hardware runs the embedding model. Requires app restart to take effect.
Run this after saving, then restart the app.
[ model suggestions ]
OpenAI
gpt-4o · gpt-4-turbo · gpt-3.5-turbo
Anthropic
claude-opus-4-7 · claude-sonnet-4-6 · claude-haiku-4-5-20251001
Ollama
mistral-small3.1 · llama3.1 · gemma2
Groq (generic)
Base URL: https://api.groq.com/openai/v1/chat/completions
llama-3.1-70b-versatile
[ security ]
API keys are encrypted with Fernet symmetric encryption before being stored in the database.
Set ENCRYPTION_KEY in your .env to a stable 44-char Fernet key. Generate with: 
python -c "from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())"