Vue d’ensemble
MediaPipe connects model inference with application inputs and interactions. Tasks, model assets, Model Maker, and Studio cover different parts of development. Platform support must be checked per task; older Legacy Solutions are distinct from current Tasks.
Fonctionnalités clés
- Task APIs
- Model assets
- Multimodal tasks
- Model Maker
- Studio
Prérequis, installation et démarrage rapide
Utilisation
How it works
Task APIs process device inputs, run models, and return structured results for the application. Supported models can be customized with Model Maker.
Audience and requirements
Mobile, web, and edge developers. Platform SDK, supported model assets, and input devices.
Practical use cases
Camera interaction; local classification; mobile ML prototypes.
Limitations and selection
Task APIs and modes differ by platform. Device performance and power consumption require measurement. Avoid assuming legacy support continues.
Related projects and selection
opencv/opencv：Complement: OpenCV handles image processing; connect formats to MediaPipe task APIs.
lutzroeder/netron：Complement: inspect supported model formats and input/output structure before integration.
Source review
Editorial analysis of upstream sources, without runtime or benchmark testing. Proposed workflows are editorial suggestions.
Compatibilité des modèles et cas d’usage
Models must match the selected Tasks interface and metadata requirements; support varies by platform.
Licence et notes sur les risques
The repository page identifies Apache-2.0. Read LICENSE; model weights and datasets may have separate terms.
Editorial source review 2026-09-09T05:00:00.950Z. README and live repository page verified; current stars/forks from GitHub HTML. Last-push metadata retained from 2026-09-05 discovery snapshot. No runtime benchmark. Integration proposals are editorial analysis.
Publication et maintenance
Reviewed 2026-09-09. Counters come from repository pages; features are based on upstream documentation. See Releases in the source links. Editorial analysis of upstream sources, without runtime or benchmark testing. Proposed workflows are editorial suggestions.