Prodigy

Prodigy is a downloadable, extensible annotation tool designed for creating training and evaluation data for machine learning models. It supports NLP and computer vision tasks including named entity recognition, text classification, object detection, and image segmentation. Prodigy features a self-hosted backend, active learning-powered models that update as you annotate, and a customizable web application interface. It enables efficient human-in-the-loop machine teaching with no machine learning experience required, full data privacy by running locally or on-premise, and tight integration with popular ML libraries like spaCy, Hugging Face, PyTorch, and TensorFlow.

platform:web platform:linux platform:macos platform:windows platform:python form:desktop-app form:web-app form:cli-tool form:library form:api form:saas pricing:paid feature:active-learning feature:custom-workflows feature:data-privacy feature:human-in-the-loop feature:local-deployment feature:custom-annotation-interfaces feature:python-sdk integration:spacy integration:huggingface integration:pytorch integration:tensorflow target:developers target:data-scientists target:ml-engineers use-case:training-data-creation use-case:machine-learning use-case:natural-language-processing use-case:computer-vision