I am trying to create a custom dx component using AI from app studio by clicking on the button “Generate Component with AI”, but i couldn’t create one. Please find attached the screenshot from where i’m trying to create it.
I couldn’t find any documents that has steps to create a DX component using AI. Could someone please help me on this?
@DineshKumarD9112 The SME’s have confirmed with me that there’s no restriction in Community Edition around this feature.
We are not sure why you might be running into this problem.
The Community Edition instances where users were given DX component code from AI all seem to run normally for other users. Are you encountering any other connectivity issues or errors in the logs?
The only “restriction” we have in our GenAI implementation are daily and monthly quotas, but your screenshot indicates that you ae not hitting either of those as we do not see an error connecting to AI services.
Hopefully another user on this forum might be able to compare notes with you.
For the benefit of other users who can comment on this thread, please show the full steps carried out and whether other functionality is working as documented.
@MarijeSchillern It also does not work for me. Select component dropdown is empty, after typing description of the component in the prompt text field i am presented with text description not code.
PS. This feature is not documented. It would be great if Pega would provide sample prompts
Something about getting valid Certs. Other suggested steps:
Open the application definition rule in dev studio, click Save button and capture the pegarules logs for exception details.
If the custom front end is correctly built using the DX API responses, it can automatically adjust to changes in Pega Platform. As a result, a single model can supply an interface that is rendered through standard Pega UI, mobile, web embed, or a custom front end.
@Pawel Sobolewski @DineshKumarD9112 This feature is demo ware, still we are working on making it robust, kindly do not use for production purpose.
We are disabling this feature in Infinity 24.1 release and will make it available once we solve all outstanding issues.
For above issue, “coming text instead of code”:
The chatgpt model text-davinci sometimes end-up producing text instead of code so if you frame prompt by giving more explicit text, it might produce code.