Reclassify the text content/image

We are quite new to OpenAI API.

We had a classifieds/ads website, and instead of creating our own AI model, we plan to utilise OpenAI to detect the listing ads that violate our terms of use, such as illegal/prescription drugs and adult content. The output will be OK or NOT OK, with a sub-output such as ADULT RELATED or DRUGS RELATED.

As a starting point, it may start with the text content first and then eventually incorporate the image later.

We’ve seen this with DevDay recently, which seems possible. Is that correct?

Also, is there any usage calculator to estimate the API cost? The average listing is 2000/week. The average of listing words is 500 words.

I appreciate your feedback.

Thank you.