Free AI Image Captioner
Turn anime, illustrations and AI artwork into comma-separated booru tags. Choose an image, describe it with local AI, then edit or copy the tags for your next prompt. No account or tokens needed.
Describe Image
PNG, JPEG, WebP and other browser-supported images, up to 50 MB and 40 megapixels. Animated images use one frame.
Device requirements & first download
Recommended: a computer with 8 GB RAM and at least 1 GB free browser storage. The first download is about 400 MB. Your browser caches model files when storage is available.
GPU processing is preferred. A slower CPU fallback runs locally when GPU processing is unavailable. Phones may be slow or run out of memory. Keep this tab open while it works.
Your selected image and result are saved locally in this browser.
Your result stays available.
Choose an image, then press Describe Image.
From an image to a useful prompt
The captioner uses AI to recognize visual features: subject count, hair and eye colors, clothing, poses and background details. It works best with anime, manga and illustrations. A typical result looks like solo, red hair, outdoors, sunlight.
- Choose, drop or paste an image. It stays on your device.
- Press Describe Image. The first run downloads the model; later runs can use the browser cache.
- Review the tags and remove anything that does not match. Copy them, use them in the generator, or turn them into a paragraph with the prompt enhancer.
The AI works with predictions, so the tool may miss details or identify them incorrectly. It does not recover the exact original generation prompt or metadata.
Image captioner questions
Are my images uploaded?
No. The image is decoded and processed on your device in a browser worker. The page downloads model and runtime files, but does not upload your image. Your selected image is temporarily stored in this browser so it can be restored after the page reloads.
Does the captioner write sentences or booru tags?
It produces English booru tags separated by commas, for example solo, red hair, outdoors, sunlight. The AI is an image tagger rather than a natural-language captioning model. You can send the tags to the prompt enhancer to turn them into a descriptive paragraph.
What device do I need?
A computer with 8 GB RAM and at least 1 GB free browser storage is recommended. The first download is about 400 MB, including the model and runtime. GPU acceleration is preferred; a slower CPU fallback runs locally when GPU processing is unavailable. Phones may be slow or run out of memory.
Can the tags recover an image's original prompt?
No. The tagger predicts visible subjects and features from pixels. It cannot recover the exact original prompt, seed or generation settings. It can miss details or return incorrect tags, so review the output before using it.
More free image tools
After tagging an image, you can generate a new image from the prompt, upscale it with AI, edit it with layers and brushes, or remove its background.