QVAC: Building local-first, peer-to-peer AI applications and systems

QVAC is an open-source framework by tetherto for building AI applications that run locally (no cloud dependency) and communicate directly between peers (peer-to-peer, or P2P). It's part of the growing "local-first" movement that pushes back against the dominant cloud-centric model where user data and AI inference happen on remote servers. Key concepts: local-first means data and processing stay on the user's own device; P2P means nodes talk to each other without a central server; and AI systems here refers to LLM-based tools and agents. This matters because it addresses privacy, cost, latency, and offline resilience — all weaknesses of current cloud-dependent AI services like ChatGPT or Claude. The project likely builds on technologies like BitTorrent-style networking or IPFS for P2P, and local models via llama.cpp or ONNX runtime, though the repo is early-stage.