Describe an Image Input

Describe an Image Input

Describe an image supplied by URL using a multimodal Responses API call.

1. Create a stored response with a multimodal input that pairs the text instruction with an input_image content part.
2. Poll the response until image understanding finishes and it leaves the in_progress status.
3. Retrieve the settled response and extract the description text and token usage, including the image URL in the returned items.

What you'll need to have ready

• apiKey: OpenAI API key used as the Bearer credential.
• model: Vision-capable model ID (e.g. gpt-4o).
• imageUrl: The URL of the image to describe.
• instruction: The text instruction guiding the description.

For your developer

Arazzo workflow chatgpt:image-input-describe, 3 operations: createResponse → getResponse → getResponse. Specs on APIs.io.