edge_veda 1.0.0

On-device LLM inference SDK for Flutter. Run Llama, Phi, and other language models locally with Metal GPU acceleration on iOS devices.

Changelog

All notable changes to the Edge Veda Flutter SDK will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Planned

• Flutter Web support (WASM + WebGPU)
• Speech-to-Text integration (Whisper)
• Text-to-Speech integration (Kokoro-82M)
• Voice Activity Detection (VAD)
• Prompt caching
• LoRA adapter support
• macOS and Windows desktop support

1.0.0 - 2026-02-04

Added

• Initial release of Edge Veda Flutter SDK
• On-device LLM inference with llama.cpp integration
• FFI bindings for native C++ core
• Hardware acceleration support (Metal on iOS, Vulkan on Android)
• Streaming text generation with real-time tokens
• Model download and management system
• Checksum verification for model integrity
• Memory-safe operations with configurable limits
• Pre-configured model registry (Llama 3.2, Phi 3.5, Gemma 2, TinyLlama)
• Comprehensive error handling with typed exceptions
• Example chat application
• Full API documentation

Features

• EdgeVeda class for LLM inference
• ModelManager for downloading and caching models
• GenerateOptions for fine-grained control
• Token streaming with TokenChunk
• Progress tracking for model downloads
• Memory usage monitoring
• Support for system prompts and JSON mode
• Stop sequences for controlled generation

Platform Support

• iOS 13.0+ with Metal acceleration
• Android API 24+ with Vulkan support

Documentation

• Complete README with quick start guide
• API reference documentation
• Example application
• Best practices and troubleshooting