An immutable persistent rope data structure
Persistent immutable array
Ropes in Rust
Compact buffer/string type for zero-copy parsing
High-performance cross-platform text engine for massive files.
Ultra low latency transformer inference with mincut-gated coherence control
Custom CUDA kernels and decode runner for Ferrum inference
A3S Power — Privacy-preserving LLM inference for TEE environments
Browser-resident Gemma 4 inference: pure Rust → WebAssembly + WebGPU. Loads Ollama's on-disk GGUF blobs and runs the forward pass on the local GPU via hand-written WGSL.
Local LoRA fine-tuning for the rullama Rust runtime, built on rullama's wgpu kernels. Compiles for native and wasm32; browser harness lives in examples/web.
OpenAI-compatible HTTP API server for OxiLLaMa
CPU backend for RLX — SIMD kernels, BLAS dispatch, thread pool, arena executor