Adding a picture questions feature

Contribute Feature

AI-generated summary

The discussion revolves around adding a picture questions feature to the Discourse AI plugin. kuaza uses a similar feature on their own site, where an uploaded image with a question is sent to ChatGPT for text parsing and response. They suggest using Google Cloud for text conversion and propose making this feature available for specific categories, allowing for interactive Q&A topics.

Alia23 agrees that this feature would be more efficient with AI integration and inquires about the plugin used for image-to-text conversion. kuaza responds that they do not use webhooks and lack the skills to code plugins. Alia23 expresses interest in implementing this feature and asks for a guide on creating it using webhooks.

1

I use the picture questions feature on my own site. It would be great to have this in the discourse ai plugin.

After uploading the image with a question in a conversation initiated with a private message, it is parsed and sent to chatgpt in text and the value is given to the user.

Google cloud can be used to convert the text on the image to text (like I am doing now).

I think it’s an idea that can be developed. What are you thinking?

2 my second thought - We can consider using this interaction with PM as a regular thread! This feature can be turned on or off for each category, so that interactive question and answer topics can be created for topics or questions opened in the desired categories, and since these topics will be open to everyone, they will appear active and affect the quality of the content in a good way. What are you thinking ?

1 Like
2

It will be even more efficient with artificial intelligence integration. Which plugin do you use to convert the expressions in the image into text? Is there a plugin for this? @kuaza

1 Like
3

Sorry for replying late. I cannot use the webhooks feature. Unfortunately, I’m not skilled enough to code plugins :slight_smile:

1 Like
4

Thanks for your answer. I want to do this too. How can I do it. Is there a guide to making it from webhooks?