Chat

Chat with a local AI model that answers questions about MishMash based on the site's content. The model runs entirely in your browser via WebLLM — no data is sent to any server. The model is downloaded once and cached locally.

Loading knowledge base…

Model: SmolLM2 1.7B (~1.0 GB) — fast Llama 3.2 1B (~0.9 GB) — fastest Llama 3.2 3B (~2.0 GB) — good balance Phi-3.5 Mini (~2.2 GB) — higher quality Gemma 2 2B (~1.6 GB)