ruvnet/ruv-fann

ruv-FANN is a comprehensive neural intelligence framework written in Rust that reimagines how artificial intelligence is built and deployed. It consists of three core components: ruv-FANN Core (a memory-safe rewrite of the classic FANN library), Neuro-Divergent (advanced forecasting models with 27+ architectures), and ruv-swarm (a distributed swarm intelligence system that achieves 84.8% SWE-Bench solve rate). The framework is designed for developers who need fast, composable, and ephemeral neural networks that run anywhere from browsers to embedded systems without requiring GPUs, making it ideal for CPU-native AI applications and edge computing scenarios.