Positron AI questions

[gvs]

Hi,
I am looking into the possibility to implement Positron AI for our documentation team, but have a few questions. We have the Enterprise edition of Oxygen Editor.

• The description of Positron AI for Enterprise indicates that it is free for Oxygen Editor Enterprise users, but there is also a mention of a 12 month description for a price and the need for a subscription and license key - Is the use and subscription/license key free or not?
  When linking it to MS Azure OpenAI the description mentions that I need an Endpoint URL, Key and the deployment. I presume (and correct me if I am wrong) that I can get this from the MS Azure admin in our company. Am I right?
  Did I forget to ask some other important questions, such as how many credits are generally used, etc.

Sorry for asking very basic questions, but I want to make sure that all bases are covered, before our start.

[Radu]

Hi,

The description of Positron AI for Enterprise indicates that it is free for Oxygen Editor Enterprise users, but there is also a mention of a 12 month description for a price and the need for a subscription and license key - Is the use and subscription/license key free or not?

If you have an enterprise version of Oxygen, the enterprise version of the AI positron add-on is free to use.
https://blog.oxygenxml.com/topics/ai_po ... __overview

When linking it to MS Azure OpenAI the description mentions that I need an Endpoint URL, Key and the deployment. I presume (and correct me if I am wrong) that I can get this from the MS Azure admin in our company. Am I right?

Yes, this is correct, with the enterprise add-on you take care of providing the AI engine to which the add-on connects.
Here's for example how I've set up the add-on on my side:

Screenshot 2024-07-09 at 08.05.09.png

We have a Microsoft Azure account which has a deployment based on gpt-35 which is a cheaper model. And we have another which has a deployment based on gpt-4 with vision support which is used by the add-on for interpreting content from inside images for example. So the add-on knows to switch to the more expensive model when it needs to read content from inside images (for example if you use our actions to create alternate text for images).

Did I forget to ask some other important questions, such as how many credits are generally used, etc.

It depends on how extensively you are using it, how much content you give it to process, for example you may use our batch AI Refactoring actions to apply the AI on all topics of your user's guide. I think that in general the price you will pay to Microsoft Azure for hosting the AI engine is low, maybe 10 USD per person.

Regards,
Radu

[gvs]

Thanks Radu,
Your answers are very clear and helpful.
Greetings
Gideon