Correct licensing model for having multiple annotators each with a model in the loop


Within a large org we wish to create a process where we spin up multiple annotation tasks where each annotator is focused on a single task with a model in the loop. Does your licensing model allow this?

Thanks for the help,

1 Like

Yes, that’s generally no problem. Prodigy is licensed per developer, which is the person who has access to the Prodigy package, back-end, CLI and library. Annotators who will only have access to the front-end don’t require a license. The only restriction is that the web app can’t be publicly accessible by everyone, and should be hosted on an internal network or behind a password-protected URL.

So if your company holds one pack of five seats, up to five developers can start Prodigy instances and tasks on your machines at the same time, run training experiments, develop custom recipes, export data etc. The company seats are floating, so if a team member leaves and someone else takes over, they can also take over the Prodigy developer license.

Hi Ines,

Following on from James question, we have a sentence annotator working well locally based on the sentiment trainer example (customer feedback sentiment). We would like to have multiple annotators working on the sentence annotations through a url front end.

I assume that we would alter the http output in the recipe from the local server, to the url where we would like our annotators to use?

Am I right in thinking that the developer would then re-send "job" to the url as required?

Many thanks