Admin · LLM Chat

🔐 Admin access

Loading…

🔌 Connection

Ollama server URL All browser calls go straight to this server (CORS must allow this origin).

🎛️ Generation parameters

Temperature 0.7 Higher = more creative / random; lower = more focused and deterministic. Top-p 0.9 Nucleus sampling: only the most likely tokens covering this probability mass. Max tokens per reply

📜 Rules for the model

Base system prompt

Additional rules (all enabled rules are appended to the system prompt)

🧠 Models

Default model used in chat

Pull a new model from the registry

Installed models

Loading…

⚠️ Disclaimers

Show the “Disclaimers” link in chat Disclaimer text (shown in the chat modal)

🧪 Advanced

Context budget (rough tokens sent per request) Older messages beyond this budget are trimmed automatically.