1. Who is behind Prodigy? Prodigy is developed by Explosion. We’re also the makers of spaCy, a leading open-source library for Natural Language Processing in Python, and several other open-source tools.

2. How long has Prodigy been around? The private beta started in summer 2017 and Prodigy has been on general sale since December 2017.

Ordering Prodigy

1. How much does Prodigy cost? Do I have to buy a subscription? Prodigy is available as a downloadable tool for a one-time fee with 12 months of free upgrades. You can choose between two different license types: a personal license or a number of transferrable company seats. See this page for more details and pricing. For the license agreement, see our terms and conditions.

2. How can I add 12 more months of upgrades to my existing license? The easiest way is via our online shop. Just click on the “Already have a license?” link under the license type you’re interested in. Make sure to enter your order ID (starting with #EX) and complete checkout. You won’t have to do anything else – your download link will stay valid. If your link has been deactivated already, please allow up to 24 hours for your access to be restored.

3. Do you have special offers for researchers and universities? Researchers at degree-granting academic institutions can apply for an interim license to use Prodigy for free in their research.

4. What does the added VAT mean? Our company is based in Berlin, Germany, so if you’re ordering from within the EU or from the UK, you will be charged an additional VAT. If you’re a VAT registered business in the UK or the EU and not based in Germany, the reverse charge mechanism applies, meaning no VAT will be added on checkout. Businesses and individuals outside the EU and UK won’t be charged an additional tax.

5. The online store tells me that my card is not valid or my ZIP code doesn’t match. What can I do? Credit card payments are processed and validated by Stripe. The ZIP code error typically occurs if the ZIP code of the billing address doesn’t match the ZIP code of the credit card, which is an additional security check we have enabled. If it’s still not working, send us an email.

6. Can I send you a purchase order and receive an invoice? Sure! For company licenses we support purchases via purchase order. We’re also already set up with most of the popular software resellers. Just email us at contact@explosion.ai.

7. Do you offer a trial version? Prodigy runs entirely on your own hardware and never phones home or connects to our servers. So we typically do trials by hosting a VM that you can log in to. This gives you the full experience of the tool, including the scriptable back-end, and also makes it easy for us to log in and help if you get stuck. If you’re interested, get in touch! Please note that we’re only able to offer VM trials to companies and organizations, not individuals.

8. Why do you not offer Prodigy for organizations that are primarily engaged in work relating to military, law enforcement or intelligence? Artificial Intelligence is an emerging technology, for which international standards and regulations have not yet been agreed. This leaves software providers to implement their own restrictions and policies. While there are many use cases of AI that can be potentially harmful, commercial projects that cause harms will be subject to ordinary civil and criminal sanctions. The same cannot be said for harmful governmental projects. Explosion has therefore decided to refuse provision of our software to public or private entities primarily engaged in work relating to military, law enforcement, intelligence and national security purposes. Regulatory agencies and tax authorities are specifically exempt (i.e., software can be provided to agencies such as the U.S. EPA, IRS or SEC). We regret that this policy may prevent uses of our software that are legally and ethically unproblematic.

Using and extending Prodigy

1. Do I need to know Python to use Prodigy? Realistically, yes: you don’t necessarily need to use the programmatic interface often, but we do assume knowledge of Python in the documentation, and if you run into problems, we’d find it difficult to help if we can’t assume a common language. Of course, if you’re just using the annotation interface, and a colleague is operating the backend, no Python knowledge is required.

2. Can I use Prodigy with other machine learning libraries like PyTorch or TensorFlow? Yes! Prodigy follows a component-based architecture and is built around recipes – simple Python scripts that control the processing logic and text source, and specify how the tool interacts with your model. The built-in NLP recipes are powered by spaCy – but you can easily write your own, using any framework or library of your choice. If you can load it in Python, you can use it with Prodigy. For a more advanced example, check out this tutorial that shows how to integrate Prodigy with TensorFlow’s object detection API.

3. Can I extend Prodigy with more interfaces? Yes! Prodigy lets you write custom interfaces using HTML, CSS and JavaScript, and combine different interface components to build fully custom UIs. The source of the REST API is included and documented, too, so you can even plug in your own front-end.

4. How can I get help and support with using Prodigy? Prodigy comes with extensive documentation, including API specs of the individual components, examples of custom recipes and data formats, as well as various tips and tricks for achieving the best results. We’re continuously working on providing more, freely available resources and educational materials like usage workflows, video tutorials and open-source scripts. We also answer user questions on the Prodigy Support Forum.

Scaling up Prodigy and “Prodigy Teams”

1. How can I use Prodigy for large annotation teams? Most annotation projects need to start with relatively few annotators, to make sure the annotation scheme and onboarding process allows high inter-annotator consistency. Once you have your annotation process running smoothly, there are a few options for scaling up your project for more annotators. One option we recommend is to divide up the annotation work so that each annotator only needs to deal with a small part of the annotation scheme. For instance, if you’re working with many labels, you would start a number of different Prodigy services, each specifying a different label, and each advertising to a different URL. Prodigy can be easily run under automation, for instance within a Kubernetes cluster, to make this approach more manageable. If you do want to have multiple annotators working on one feed, Prodigy has support for that as well via named multi-user sessions. You can create annotator-specific queues using query parameters, or use the query parameters to distinguish the work of different annotators so you can run inter-annotator consistency checks.

2. When are you launching Prodigy Teams? Where can I sign up? Prodigy Teams is our long-awaited product, that takes care of all the tasks mentioned above for working collaboratively on annotation across an organization. As of September 2023, we’ve had successful Beta testing rounds with a handful of testers and quickly iterating on features to get it ready for launch. We have written a Roadmap that includes more details and how Prodigy Teams fits in with several planned updates to Prodigy as well.
   
   Sign up for the waitlist or beta testing