MediaPipe

MediaPipe is an open-source toolkit for developers adding on-device AI to applications on Android, iOS, the web, desktop and edge devices. It pairs pretrained models with APIs for specific tasks, so developers can use existing solutions or customize them for their applications. The project uses the Apache 2.0 license.

Its vision tools cover face, hand and pose landmarks, gesture recognition, object detection and image segmentation. Image classification and embeddings support applications that need to categorize or compare images. Audio classification sits alongside text tools for language detection, classification, embeddings, proofreading and summarization. An LLM Inference API is also available. Platform support varies by solution.

MediaPipe Tasks provides the application-facing libraries, with Python support alongside mobile and web APIs. Model Maker lets developers adapt supported models using their own data. MediaPipe Studio runs in a browser and provides tools to visualize results, evaluate solutions and benchmark performance. For developers building custom processing pipelines, MediaPipe Framework provides a lower-level foundation for on-device machine learning.

MediaPipe Tasks processes input such as images, video and text on the device and doesn't send that input to Google servers. The APIs do send performance and usage metrics to Google, a distinction that matters for applications handling data that must stay local.

Some legacy solutions are unsupported and provided as-is. Use the current task documentation to check platform availability and model-customization support.