llamacpp library

LlamaCPP backend for RunAnywhere Flutter SDK.

This module provides LLM (Language Model) capabilities via llama.cpp. It is a thin wrapper that registers the C++ backend with the service registry.

Architecture (matches Swift/Kotlin)

The C++ backend (RABackendLlamaCPP) handles all business logic:

  • Service provider registration
  • Model loading and inference
  • Streaming generation

This Dart module just:

  1. Calls rac_backend_llamacpp_register() to register the backend
  2. The core SDK handles all LLM operations via rac_llm_component_*

Quick Start

import 'package:runanywhere_llamacpp/runanywhere_llamacpp.dart';

// Register the module (matches Swift: LlamaCPP.register())
LlamaCpp.register();

// Register models through RunAnywhere.models.
// The commons registry/router owns framework selection and routing.

Classes

LlamaCpp
LlamaCPP module for LLM text generation.