> ## Documentation Index
> Fetch the complete documentation index at: https://docs.counso.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Ask an Agent to inspect an image

Image attachments let an Agent analyze visual material in the context of the task and any knowledge it can access. This is useful for reading a screenshot, comparing a design with brand guidance, checking a product image, or identifying visible differences between documents. Image understanding depends on the model selected for the Agent and the quality of the file.

Attach a clear image in the conversation. State what the Agent should inspect and which reference source to use. For a brand review, provide the current brand guidelines and ask for each finding to cite the relevant visual element and guidance. For a receipt or scanned document, ask it to transcribe legible fields and flag anything uncertain instead of filling gaps.

A good review separates observation from interpretation: describe the visible color, text, alignment, or object first, then explain the comparison. Ask the Agent to identify unreadable areas and avoid inferring details that are not visible. Verify important measurements, text, and compliance conclusions against the original image or source document before acting.

If the Agent cannot inspect the image, confirm that the chosen model supports visual input and that the file is within the upload limit. A low-resolution, cropped, rotated, or blurry image may need a clearer copy.