TechnologiesArea 8

Machine Learning & AI

28 frameworks

The Machine Learning & AI area provides the frameworks for adding intelligence to your app, running models on device to protect user privacy and reduce latency. With these frameworks you integrate trained Core ML models, prompt on-device large language models for structured generation, and apply task-specific capabilities such as computer vision, natural language processing, speech recognition, and sound analysis. You can also train and evaluate custom models, generate images and translations, and accelerate neural-network workloads on the GPU.

iOSmacOStvOSwatchOSvisionOS
No frameworks match.
CoreML63 types
Run and integrate trained machine learning models on device, providing features, batch prediction, and compute-device control.
iOSmacOStvOSwatchOSvisionOS
FoundationModels42 types
On-device large language model APIs: prompt, generate structured @Generable content, and run guided generation sessions.
iOSmacOStvOSwatchOSvisionOS
Vision202 types
Performs computer-vision analysis on images and video, including detection, recognition, body/animal pose, and aesthetics.
iOSmacOStvOSwatchOSvisionOS
NaturalLanguage16 types
On-device natural language processing: tokenization, language ID, word and contextual embeddings, tagging and sentiment.
iOSmacOStvOSwatchOSvisionOS
CoreAI
Framework providing delegation interfaces for on-device AI and machine learning operations across Apple platforms.
iOSmacOStvOSwatchOSvisionOS
CreateML49 types
Trains and evaluates custom Core ML models (classifiers, regressors, image/text/sound/tabular) on device.
iOSmacOStvOSwatchOSvisionOS
CreateMLComponents176 types
Composable building blocks for building, training, and evaluating custom machine learning models and feature pipelines.
iOSmacOStvOSwatchOSvisionOS
DataDetection10 types
Detect and extract common data types (links, phone numbers, email addresses, dates) from text via DDMatch result objects.
iOSmacOStvOSwatchOSvisionOS
ImagePlayground6 types
Generates images programmatically from text descriptions and style information using the system's generative models.
iOSmacOStvOSwatchOSvisionOS
LatentSemanticMapping3 types
Carbon-era latent semantic analysis for classifying and categorizing text by meaning.
iOSmacOStvOSwatchOSvisionOS
MediaIntelligence9 types
Face detection, clustering, and recognition framework using CoreData persistence and machine learning to group and identify individuals across images.
iOSmacOStvOSwatchOSvisionOS
MetalPerformanceShadersGraph51 types
A graph framework for compiling and running GPU-accelerated compute and neural-network operations via Metal.
iOSmacOStvOSwatchOSvisionOS
MLCompute75 types
Low-level GPU/CPU-accelerated neural network primitives (layers, optimizers, graphs) for training and inference.
iOSmacOStvOSwatchOSvisionOS
MusicUnderstanding9 types
Machine learning framework for analyzing musical properties including instrument activity, loudness, and other acoustic characteristics.
iOSmacOStvOSwatchOSvisionOS
SensitiveContentAnalysis4 types
Analyzes images and video streams on-device to detect sensitive or explicit content and report sensitivity results.
iOSmacOStvOSwatchOSvisionOS
ShazamKit14 types
Recognizes audio by matching it against Shazam's catalog or custom reference signatures, returning matched media metadata.
iOSmacOStvOSwatchOSvisionOS
SiriInferenceLearning
Framework for Siri inference and machine learning operations across Apple platforms.
iOSmacOStvOSwatchOSvisionOS
SoundAnalysis11 types
Analyze streamed or file audio with Core ML sound classification models to detect and classify sounds over time.
iOSmacOStvOSwatchOSvisionOS
Speech36 types
Perform speech recognition, transcribing audio files and live audio with support for custom language models.
iOSmacOStvOSwatchOSvisionOS
Translation3 types
Checks language-pair availability and performs on-device text translation, with SwiftUI presentation support.
iOSmacOStvOSwatchOSvisionOS
TrustInsights6 types
Framework for evaluating trust and security insights using version-managed assessment models.
iOSmacOStvOSwatchOSvisionOS
VisionKit11 types
Document scanning, text/data recognition and visual lookup using the camera, returning scanned documents and live text.
iOSmacOStvOSwatchOSvisionOS
VisualIntelligence1 types
Lets apps surface relevant content from the Visual Intelligence scene via App Intents, using the captured pixel buffer.
iOSmacOStvOSwatchOSvisionOS
_CoreSpotlight_FoundationModels overlay17 types
Integration layer enabling foundation models to search and reason over app-indexed content via CoreSpotlight sources.
iOSmacOStvOSwatchOSvisionOS
_FoundationModels_AppKit overlay
AppKit convenience initializers for FoundationModels Attachment types, enabling NSImage integration with AI model operations.
iOSmacOStvOSwatchOSvisionOS
_FoundationModels_SwiftUI overlay
Cross-import overlay bridging FoundationModels with SwiftUI to integrate on-device machine learning models into SwiftUI applications.
iOSmacOStvOSwatchOSvisionOS
_Translation_SwiftUI overlay
Cross-import overlay adding SwiftUI view modifiers to present the system translation popover for on-device text translation.
iOSmacOStvOSwatchOSvisionOS
_Vision_FoundationModels overlay2 types
Cross-import overlay bridging Vision with FoundationModels to provide barcode-decoding tools for language model image analysis.
iOSmacOStvOSwatchOSvisionOS
← All technologies