{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/8d0d0c5253794399ac75ef62a092622b\" frameborder=\"0\" width=\"1548\" height=\"1161\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1161,"width":1548,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1161,"thumbnail_width":1548,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/8d0d0c5253794399ac75ef62a092622b-f6ad0c4a3c6f53fb.gif","duration":299.947,"title":"Paw Paw Plus AI Pet Care Planner","description":"In my Loom I present Paw-Pow Plus, an AI powered pet care planning system. It schedules daily tasks, then uses a RAG augmented agent to resolve conflicts and apply pet specific knowledge with hard safety guarantees baked into Python tool executors. The stack uses Grox and a Llama 3 370B model, plus TF IDF retrieval from a 39 chunk knowledge base for topics like arthritis, anxiety, and medication. The agent has five tools, reschedule, add tasks, root tasks, search knowledge, and mark the result. I include demos in my GitHub for three scenarios, and I share the reliability and guardrail approach. I do not request any action from viewers."}