Creative Technology · Computer Vision

AI-Powered Image Captioning

Choose an example or upload your own image. Then choose a pretrained captioner and let AI answer: what is happening in this image?

Task Image captioning Output A description in English Models ViT-GPT2 · Florence-2 · SmolVLM

Choose an image

Start with an example, or try a photo of your own.

Start here · Example images

Four images from a bank of 16 teaching examples.

or try your own

Upload an image

PNG, JPEG, or WebP · up to 12 MB.

Choose captioner

Description style

Maximum new tokens64

An output limit, not an exact word count. A low limit may cut a sentence short.

Choose an example or upload your own image.

AI description

How does the model describe it?

Waiting

Choose an image first. It will appear here before captioning.

No caption yet.

Check the description	Look for visual evidence
Objects and attributes	Are the named objects and their details visible?
Actions and relationships	Does the sentence explain what is actually happening?
Missing or added details	What is omitted? What has the model added without evidence?

What you are doing

Use a learned AI capability instead of rebuilding the traditional task from scratch.

A visual encoder represents the image. A language decoder generates a description. Different models may describe different details, and a fluent sentence can still be wrong.

Example or your image→Selected pretrained captioner→A description in words

Compare models: keep the image fixed, switch captioner, and check the objects, actions, and relationships in each sentence.

Comparison log

Keep one image fixed and compare descriptions.

The latest 12 successful runs stay here until you clear the log or reload. Generation time excludes model loading; download the log to keep the full settings.

Run	Image	Model	Caption	Style	Time
Run two models on the same image to start a comparison.