Enter your custom prompt/instruction in the top box. Add any context, data, or input in the bottom box. Click Generate to send it to the model and see the raw output.