You're in demo mode
Back to site

Settings

Configure model, parameters, status, and white-label.

Basic information
Name and description visible to your team.
Model & sampling
Choose the underlying LLM and inference parameters.
0.20

Lower = more deterministic. Receptionist-style agents work best at 0.2–0.4.

RAG retrieval
How many chunks the agent retrieves per query, and the minimum similarity to count as relevant.
5
0.40
Visibility & branding

When on, the embed widget is reachable without authentication.

Fully white-label this agent's widget — no mention of us.

Demo mode — your changes won't persist.