Question

A developer is building a feature that lets users upload a photo of a handwritten recipe so Claude can transcribe it. They are new to the Messages API and want to understand, at a conceptual level, how to send the image to Claude. Which statement correctly describes how image inputs work in the Messages API?

Accepted Answer

An image is included as a content block within a user message, alongside any text content blocks, so Claude can analyze the picture and the accompanying instructions together.

Answer

Images must be placed in the system prompt because that is the only place the model reads non-text data.

Answer

The image must be sent as a separate assistant-role message before the user asks their question.

Answer

Claude cannot process images directly; the developer must run OCR themselves and paste the extracted text into a user message.