Booru | Caption

Want to contribute? Here is the standard workflow for creating a caption that will get "favorited."

: Models trained predominantly on realistic photography (like standard FLUX or SDXL) do not understand booru tagging formats well. Feeding them comma-separated tags instead of natural sentences can result in broken anatomy or low-quality generations.

If you are interested in exploring a Caption Booru (such as the current active versions of captionbooru.com or the various .art successors), follow these guidelines: Caption Booru

Before diving into the niche of "Caption Booru," it's essential to understand the term "Booru" itself. A booru (a phonetic rendering of "board") is a type of tag-based imageboard or image gallery originally designed for sharing and archiving images, typically anime-style illustrations. Unlike standard social media or image hosting, Boorus rely heavily on a sophisticated tagging system. Each image is categorized by tags that describe nearly everything visible within it, from characters to poses to specific background elements. This tag-centric approach shifts the focus from a title or description to data-driven searchability. While platforms like Danbooru and Gelbooru are the titans of this space, they traditionally focus more on tags than on narrative descriptions. However, a subculture emerged where the narrative took center stage: the world of .

: A newer generation of vision-language models capable of outputting highly detailed booru-style structural breakdowns alongside natural sentences. Want to contribute

For the curious reader: Start with "Safe" rated filters. Search for a tag like slow_burn_tf . Find a caption that is longer than 300 words. Read it. Chances are, if you enjoy the blend of visual suggestion and written narrative, you will be hooked.

Unlike traditional galleries that organize images into single folders, Booru platforms allow users to stack dozens of metadata tags on a single file. Commas separate these tags, and spaces are replaced by underscores (e.g., blue_eyes ) to keep individual concepts unified. If you are interested in exploring a Caption

As AI models, for example, evolve toward better understanding, for example, spatial relationships and, for example, nuanced, for example, aesthetics, the, for example, demand for, for example, high-quality, for example, captioning data will increase.

Look at the associated text tags or long-form descriptions.

Because of the Booru's open nature, different users might take the same image and write entirely different captions, showcasing the breadth of human imagination. Why the Booru Format Works for Captions

: Content often focuses on specific sexual fantasies such as (female dominance), Corruption Sissification Bimbofication (netorare/cuckolding). "Rescued" or "Anti-NTR" Captions