Demo Flow: 1. Chat with Model A about a topic → make decisions, state preferences 2. Click Save to Filecoin → conversation is summarized and uploaded as JSON via Filebase 3. Switch to Model B (fresh chat, zero prior messages) 4. Click Load from Filecoin → the saved context is injected into Model B's system prompt 5. Ask a follow-up → Model B answers with full context from Model A's conversation
Model ALlama 3.1 8B (HuggingFace)
Chat with Model A. Then save the memory to Filecoin.
Model BLlama 3.3 70B (Groq)
Load context from Filecoin, then chat with Model B.