## Overview
The **RunAnywhere Flutter SDK** is a production-grade, on-device AI SDK for Flutter applications. It enables developers to run AI models directly on iOS and Android devices without requiring network connectivity for inference, ensuring minimal latency and maximum privacy for your users.
The SDK provides a unified interface to multiple AI capabilities:
Text generation with streaming support via Dart Streams
Speech-to-text transcription with Whisper models
Neural voice synthesis with Piper TTS
Function calling with RunAnywhereTools and sealed ToolValue types
Real-time voice activity detection with Silero VAD
## Key Capabilities
* **Multi-backend architecture** – Choose from LlamaCPP (GGUF models) or ONNX Runtime
* **Cross-platform** – Single codebase for iOS and Android
* **Dart-native** – Built with async/await and Streams for reactive programming
* **Production-ready** – Built-in analytics, logging, and model lifecycle management
## Core Philosophy
All AI inference runs locally, ensuring low latency and data privacy. Once models are
downloaded, no network connection is required for inference.
Backend engines are separate packages—include only what you need. This keeps your app bundle
size minimal.
Audio and text data never leaves the device unless explicitly configured. Only anonymous
analytics are collected by default.
Subscribe to SDK events for reactive UI updates and observability.
## Features
### Language Models (LLM)
* On-device text generation with streaming support
* Dart Stream-based token streaming
* System prompts and customizable generation parameters
* Support for thinking/reasoning models
* LlamaCPP backend for GGUF models
### Speech-to-Text (STT)
* Real-time streaming transcription
* Batch audio transcription
* Multi-language support
* Whisper-based models via ONNX Runtime
### Text-to-Speech (TTS)
* Neural voice synthesis with Piper TTS
* System voices via platform TTS
* Streaming audio generation for long text
* Customizable voice, pitch, rate, and volume
### Voice Activity Detection (VAD)
* Energy-based speech detection with Silero VAD
* Configurable sensitivity thresholds
* Real-time audio stream processing
### Tool Calling
* Register typed tool definitions via RunAnywhereTools
* Automatic tool execution with configurable limits
* Dart 3 sealed class ToolValue types with pattern matching
* Multi-tool chaining for complex agent workflows
### Voice Agent Pipeline
* Full VAD → STT → LLM → TTS orchestration
* Complete voice conversation flow
* Push-to-talk and hands-free modes
## System Requirements
| Platform | Minimum Version |
| -------- | --------------- |
| Flutter | 3.10.0+ |
| Dart | 3.0.0+ |
| iOS | 14.0+ |
| Android | API 24 (7.0+) |
ARM64 devices are recommended for best performance. Metal GPU acceleration on iOS and NEON SIMD on
Android provide significant speedups over CPU-only inference.
## Package Composition
| Package | Size | Purpose |
| ---------------------- | --------- | ------------------------------------ |
| `runanywhere` | \~5MB | Core SDK (required) |
| `runanywhere_llamacpp` | \~15-25MB | LLM text generation with GGUF models |
| `runanywhere_onnx` | \~50-70MB | STT/TTS/VAD via ONNX Runtime |
## Architecture
```mermaid theme={null}
graph TB
A(Your Flutter Application)
A --> B(RunAnywhere Flutter SDK)
B --> C(RunAnywhere API)
B --> D(EventBus)
B --> E(ModelRegistry)
C & D & E --> F(DartBridge - FFI to C++)
F --> G(LlamaCpp - LLM)
F --> H(ONNX - STT/TTS/VAD)
F --> I(Future Backends)
style A fill:#334155,color:#fff,stroke:#334155
style B fill:#ff6900,color:#fff,stroke:#ff6900
style C fill:#ff6900,color:#fff,stroke:#ff6900
style D fill:#475569,color:#fff,stroke:#475569
style E fill:#475569,color:#fff,stroke:#475569
style F fill:#fb2c36,color:#fff,stroke:#fb2c36
style G fill:#475569,color:#fff,stroke:#475569
style H fill:#475569,color:#fff,stroke:#475569
style I fill:#475569,color:#fff,stroke:#475569
```
## Starter Example
Complete working example with LLM chat, STT, TTS, and Voice Agent demos
## Next Steps
Add the SDK packages via pub.dev
Build your first AI feature in minutes