I had a perfectly functioning scenario using the openai “Make an Edit” module up until about a week ago or so.
It was using the divinci model, which i know is now deprecated, but I went in to choose a new model, and none of the models populate in the module.
I deleted my openai connection, created a new API key, set up a new connection, created a new model, and still nothing.
I know that it should be “gpt-3.5-turbo” so , checked the “map” button, entered that text string in the model area, saved it, turned off the “map” toggle, and now it shows as one of the models, but im still getting an error.
See screenshots:
“Map” Workaround:
Error:
New Scenario New Connection New Module:
What am i missing? This is pretty critical for my job. would hate to have to move to zap.
The recommended replacement for the /v1/edits endpoint is /v1/chat/completions, and the “Create a Completion” module uses that endpoint. The “Edits” module won’t work because the endpoint it hits is no longer supported by OpenAI. I don’t know if make will change the module, but you should be able to recreate the “edits” functionality using the “Create a Completion” module.
I just wanted to step in and confirm what has been said.
The ‘Create an Edit’ endpoint is being deprecated by OpenAI. As a result, we cannot guarantee it’ll keep working as intended. It might eventually stop working altogether.
Our dev team will still run tests to explore what can be done, but at this point, we cannot guarantee a solution.
But how would I use a text string from a previous module within the "complete a creation’ field? It doesnt seem like it lets us use variables from previous models.
