RunAnywhere class - runanywhere library

Constructors

RunAnywhere()

Properties

hashCode → int: The hash code for this object.
no setterinherited
runtimeType → Type: A representation of the runtime type of the object.
no setterinherited

Methods

noSuchMethod(Invocation invocation) → dynamic: Invoked when a nonexistent method or property is accessed.
inherited
toString() → String: A string representation of this object.
inherited

Operators

operator ==(Object other) → bool: The equality operator.
inherited

Static Properties

currentModelId → String?: Get the currently loaded LLM model ID Returns null if no LLM model is loaded.
no setter
currentSTTModelId → String?: Get the currently loaded STT model ID Returns null if no STT model is loaded.
no setter
currentTTSVoiceId → String?: Get the currently loaded TTS voice ID Returns null if no TTS voice is loaded.
no setter
currentVLMModelId → String?: Get the currently loaded VLM model ID
no setter
environment → SDKEnvironment?: Current environment
no setter
events → EventBus: Event bus for SDK events
no setter
initParams → SDKInitParams?: Get initialization parameters
no setter
isActive → bool: Check if SDK is active
no setter
isModelLoaded → bool: Check if an LLM model is currently loaded
no setter
isSDKInitialized → bool: Check if SDK is initialized
no setter
isSTTModelLoaded → bool: Check if an STT model is currently loaded
no setter
isTTSVoiceLoaded → bool: Check if a TTS voice is currently loaded
no setter
isVLMModelLoaded → bool: Check if a VLM model is currently loaded
no setter
isVoiceAgentReady → bool: Check if the voice agent is ready (all required components loaded).
no setter
serviceContainer → ServiceContainer: Access to service container
no setter
version → String: SDK version
no setter

Static Methods

askAboutImage(String question, {required VLMImage image}) → Future<String>: Ask a question about an image
availableModels() → Future<List<ModelInfo>>: Get all available models from C++ registry.
cancelGeneration() → Future<void>: Cancel ongoing generation
cancelVLMGeneration() → Future<void>: Cancel ongoing VLM generation
chat(String prompt) → Future<String>: Simple text generation - returns only the generated text
cleanupVoiceAgent() → void: Cleanup voice agent resources.
currentLLMModel() → Future<ModelInfo?>: Get the currently loaded LLM model as ModelInfo Matches Swift: RunAnywhere.currentLLMModel
currentSTTModel() → Future<ModelInfo?>: Get the currently loaded STT model as ModelInfo Matches Swift: RunAnywhere.currentSTTModel
currentTTSVoice() → Future<ModelInfo?>: Get the currently loaded TTS voice as ModelInfo Matches Swift: RunAnywhere.currentTTSVoice (TTS uses "voice" terminology)
deleteStoredModel(String modelId) → Future<void>: Delete a stored model
describeImage(VLMImage image, {String prompt = "What's in this image?"}) → Future<String>: Describe an image with a text prompt
downloadModel(String modelId) → Stream<DownloadProgress>: Download a model by ID
generate(String prompt, {LLMGenerationOptions? options}) → Future<LLMGenerationResult>: Full text generation with metrics
generateStream(String prompt, {LLMGenerationOptions? options}) → Future<LLMStreamingResult>: Streaming text generation
getCurrentEnvironment() → SDKEnvironment?: Get current environment (alias for environment getter) Matches Swift pattern for explicit method call
getDownloadedModelsWithInfo() → Future<List<StoredModel>>: Get downloaded models with their file sizes.
getStorageInfo() → Future<StorageInfo>: Get storage info including device storage, app storage, and downloaded models.
getVoiceAgentComponentStates() → VoiceAgentComponentStates: Get the current state of all voice agent components (STT, LLM, TTS).
initialize({String? apiKey, String? baseURL, SDKEnvironment environment = SDKEnvironment.development}) → Future<void>: Initialize the SDK
initializeWithParams(SDKInitParams params) → Future<void>: Initialize with params
loadModel(String modelId) → Future<void>: Load a model by ID
loadSTTModel(String modelId) → Future<void>: Load an STT model
loadTTSVoice(String voiceId) → Future<void>: Load a TTS voice
loadVLMModel(String modelId) → Future<void>: Load a VLM model by ID
processImage(VLMImage image, {required String prompt, int maxTokens = 2048, double temperature = 0.7, double topP = 0.9}) → Future<VLMResult>: Process an image with VLM
processImageStream(VLMImage image, {required String prompt, int maxTokens = 2048, double temperature = 0.7, double topP = 0.9}) → Future<VLMStreamingResult>: Stream image processing with real-time tokens
refreshDiscoveredModels() → Future<void>: Re-discover models on the filesystem via C++ registry.
registerModel({String? id, required String name, required Uri url, required InferenceFramework framework, ModelCategory modality = ModelCategory.language, ModelArtifactType? artifactType, int? memoryRequirement, bool supportsThinking = false}) → ModelInfo: Register a model with the SDK.
registerMultiFileModel({String? id, required String name, required List<ModelFileDescriptor> files, required InferenceFramework framework, ModelCategory modality = ModelCategory.embedding, int? memoryRequirement}) → ModelInfo: Register a multi-file model with the SDK.
removeModel(String modelId) → Future<void>: Remove a model from the C++ registry
reset() → Future<void>: Reset SDK state
startVoiceSession({VoiceSessionConfig config = VoiceSessionConfig.defaultConfig}) → Future<VoiceSessionHandle>: Start a voice session with audio capture, VAD, and full voice pipeline.
synthesize(String text, {double rate = 1.0, double pitch = 1.0, double volume = 1.0}) → Future<TTSResult>: Synthesize speech from text.
transcribe(Uint8List audioData) → Future<String>: Transcribe audio data to text.
transcribeWithResult(Uint8List audioData) → Future<STTResult>: Transcribe audio data with detailed result.
unloadModel() → Future<void>: Unload current model
unloadSTTModel() → Future<void>: Unload the currently loaded STT model Matches Swift: RunAnywhere.unloadSTTModel()
unloadTTSVoice() → Future<void>: Unload the currently loaded TTS voice Matches Swift: RunAnywhere.unloadTTSVoice()
unloadVLMModel() → Future<void>: Unload the currently loaded VLM model
updateModelDownloadStatus(String modelId, String? localPath) → Future<void>: Update the download status for a model in C++ registry