This extension is a reimagined version based on the a/ComfyUI-QualityOfLifeSuit_Omar92 extension, and it supports integration with ChatGPT through the new OpenAI API. NOTE: See detailed installation instructions on the a/repository.
https://github.com/omar92/ComfyUI-QualityOfLifeSuit_Omar92
I rewrote the core logic because this code was not compaitible with the new version of the OpenAi api
2//24/2024 -I have updated logic to get to the roles and the config json files to work with Linux and MacOs.
2/21/2024 -I removed the text around the OpenAi response so now the response from chatgpt will go straight into the image generator. This makes roles useless so i will eventually remove that.
1/4/2024 -I removed max_words from the node as it didnt really seem to do anything. You can try inputting it in the prompt to see if that makes a difference.
-Sometimes, for whatever reason, chat-gpt api just doesnt work so i added logic to try 3 times waiting 10 seconds in between each try. This seems to work 100% of the time
-Added a seed so that the prompt can be resubmitted to get a difference response. Make sure that the "control_after_generate" is set to random if you want this. if not you will use the same response the first time you submit a request to chat gpt
Single node to prompt ChatGPT and will return an input for your CLip TEXT Encode Prompt
ComfyUI is an advanced node-based UI that utilizes Stable Diffusion, allowing you to create customized workflows such as image post-processing or conversions.
Download the zip file. Extract to ..\ComfyUI\custom_nodes. Restart ComfyUI if it was running (reloading the web is not enough). You will find my nodes under the new group O/....
The file looks like this :
{ "openAI_API_Key": "sk-#################################" }
This node harnesses the power of chatGPT, an advanced language model that can generate detailed image descriptions from a small input.
config.json
file