About
Kotlin-first llama.cpp integration for on-device and remote LLM inference.
Features
- Kotlin Multiplatform: shared code across Android, iOS, and desktop
- Offline inference via llama.cpp (compiled with Kotlin/Native bindings)
- Remote inference via optional HTTP client (e.g. llamatik-server)
- Embeddings and text generation support
- Works with GGUF models (e.g. Mistral, Phi, LLaMA)
- Lightweight and dependency-free runtime