Perhaps the most obvious way to approach this is to define a new tool that does this:
- Reads the HTML
- Renders it using Kaleido or Choreographer into an image
- Submits the image to the LLM via the completion API
One use case is to ask the bot to incorporate the data from a web page into its response.
Perhaps the most obvious way to approach this is to define a new tool that does this:
One use case is to ask the bot to incorporate the data from a web page into its response.