Repogen

 Repogen is the open inference network. We route AI workloads to a distributed network of Apple Silicon machines running MLX-optimized models. Developers get OpenAI-compatible API access to frontier         Open-source models at roughly half the cost of centralized providers. Mac owners earn USDC from compute they already own.                                                                                     The core insight is simple: there are over 100 million Macs worldwide, most sitting idle for 16+ hours per day. Apple Silicon's unified memory architecture makes these machines exceptional for LLM inference. An M2 Ultra with 192GB unified memory can serve 70B parameter models locally. We built the coordination layer that turns this idle capacity into a permissionless inference marketplace.                                                         Repogen is not another GPU cloud. We own no hardware. We operate no data centers. We simply connect inference demand with distributed supply, handle load balancing and routing, meter usage, and settle payments in USDC on Base. The network is live today with providers earning and developers building.