Prodigy

Prodigy is an annotation tool for developers and research teams building training and evaluation datasets for custom AI models. It runs on your own machines and provides a web interface where annotators can label data without programming knowledge.

• Text and image labeling: Create annotations for named entity recognition, text classification, object detection and image segmentation, along with model evaluation tasks.
• Model-assisted review: Use model suggestions to help label examples, correct generated annotations and refine classification schemes against real data.
• Training integrations: Work with spaCy and Hugging Face models, or use the resulting annotations in custom training with PyTorch and TensorFlow.
• Custom workflows: Adapt data processing and annotation interfaces with Python, combine interface components, or build custom interfaces with HTML, CSS and JavaScript.

Prodigy is paid software that can run locally, on premises or on a cloud provider you choose. Local workflows work without an internet connection, including on air-gapped machines, and keep data and models within your environment. Optional LLM API integrations connect to external providers for model-assisted labeling, so those workflows aren't fully offline.

The developer side assumes basic Python and command-line familiarity. Annotators don't need that expertise once a task is ready, and you own the models you produce.