, and custom attributes. Right now I send the entire HTML directly to OpenAI’s API translation and get a translated response. However, when the content is too big (token limit) I don’t know what to do.

Example Input:

Welcome to our new platform. 
Consult terms and conditions for more details.

How can I efficiently split large HTML content into smaller chunks for multiple OpenAI API requests without breaking the structure?

I have tried sending only the extracted text without HTML to OpenAI API for translation, but it is not reliable and better to send whole content without any change.