Danbooru Tag Wiki Database with 640-Dim Text Embeddings

Danbooru's general-category tag wiki pages are stored in a single-file SQLite database. The database includes a sqlite-vec virtual table holding 640-dimensional embeddings of each cleaned wiki body text. JackBinary created this resource, which was last updated on Hugging Face in May 2026.

Use Cases

• Natural-language tag search based on wiki description embeddings.
• Semantic clustering of Danbooru tags based on their wiki content.
• Building a tag recommendation engine based on descriptive phrase matching.

Strengths

• Includes a sqlite-vec virtual table with 640-dimensional embeddings for each wiki body.
• Built specifically to enable natural-language search over Danbooru's tag vocabulary.
• Source code for the fetcher, embedder, and query CLI is publicly available on GitHub.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and overall dataset size are unknown, which may limit suitability assessment.
• Freshness should be verified as the last update timestamp is from the future (2026-05-21).

Provenance

Source

Danbooru tag wiki pages.

Collection Method

Fetched, cleaned, and embedded using source code from github.com/JackBinary/danbooru-db.

Freshness

Last updated 2026-05-21 14:35:27; freshness should be verified.