Hi everyone! 
Many of you have asked over the years how to scale Prodigy beyond a single developer’s machine — how to bring in more annotators, run projects as a team, and automate more of the workflow. That’s the vision we first set out to build with Prodigy Teams, and today we’re excited to share what it has grown into: Ellf, now in beta! 
Ellf is our new platform for scaling up and automating data development — built around Prodigy at its core. We've put together a page just for the Prodigy community: 
Ellf × Prodigy annotation tool · Ellf · Your virtual NLP engineer
What you can do with Ellf:
Collaborate as a team — invite developers and annotators, manage users, and monitor annotation progress in one place
Run Prodigy in your own cloud — annotation tasks are served from a data-private cluster you run locally or in your own infrastructure, so no sensitive data ever leaves your servers
Auto-annotate with agents — let LLM-powered agents handle the routine work while your team focuses on the hard cases
Design and troubleshoot your NLP solutions with an expert at your side — Ellf's assistant draws on Explosion's distilled know-how from over 10 years of applied NLP work powered with a knowledge base covering Prodigy, spaCy, Ellf and applied NLP best practices
Bring your custom recipes — port over your existing Prodigy workflows, or describe what you need and have the built-in assistant (with Claude Code integration) create new ones for you
Beyond text — annotate PDFs, images, and audio, and train models right on the platform
Data privacy and scriptability have always been core to Prodigy, and that hasn't changed — everything that touches your data runs on your own cluster.
We're currently in beta and onboarding users from the waitlist. If you'd like to try it out, you can join the waitlist or read the docs to learn more.
We'd love to hear what you think — questions and feedback very welcome in this thread! 
Thanks, as always, for all your support 