Llama Enterprise Chat

Model Router

Auto-detect loaded model and switch by (un)loading via llama-server router mode.

Available models

Loaded: Unknown

Generation

Streaming and non-streaming chat controls.

Mode

Temperature

Max tokens

Use RAG (MariaDB Vector)

RAG Store

Embeddings + MariaDB 11.8.5 (Vector) for retrieval.

DB: Unknown

Embedding model

Active: Unknown

Host

Port

User

Password

Database name

TLS

Embed Documents / Code

Uploads are chunked and embedded; vectors are stored in MariaDB for retrieval.

Collection

Files

Documents

Chat

Results are rendered as Markdown with safe sanitization.

Auto-scroll: ON

Message

RAG Hits

Debug Log

Diagnostics

Connectivity checks against llama-server endpoints.