Skip to main content

gemma3:1b · home Ollama · zero data egress

Local AI Assistant

Multi-turn chat with the home Ollama instance. Pick a system-prompt persona, ask questions, optionally analyze an image. Nothing leaves the home network — everything runs on Gary's GPU.

Persona / system prompt

Vision (image)

Welcome — pick a persona on the left, ask anything below. First reply takes ~5s.

Endpoints: POST /api/public/ai/ask · POST /api/public/vision. Conversation kept client-side only — refresh wipes it.

← Plain Ask · Narrative Engine · Models