Image Description / Captioning

A written description of the scene is produced at the length you ask for, from a short caption to a fuller archive record. An alt-text variant is generated alongside it, which makes a large image library accessible without manual work. Descriptions can be constrained to what is visible, avoiding inference about people or context that the image does not support. Output is available in several languages.

Capacity can be permanent or temporary: a one-off archive can run on burst capacity and release it afterwards, while a continuous workload sits on a cluster reserved for you. Our AI advisors help pick the model and the hardware before anything is committed.