What it does
This MCP server lets an AI client look up Danbooru tags instead of recalling them from memory. It returns exact tag names, aliases, categories, post counts, co-occurrence statistics and wiki rules. Its eight tools cover keyword search, exact and batch lookup of up to 20 names per call, alias resolution, related tags, structured wiki pages, post search with full tag lists, and distribution profiles showing what a tag typically renders alongside. Tags come back as plain strings that can be pasted into ComfyUI, Stable Diffusion or Illustrious prompt fields.
It suits local AI image generation, where a model would otherwise guess tag spellings such as bluehair instead of blue_hair. Each name is also checked against the WD14 tagger vocabularies and marked live, tagger_era or unknown, which helps when LoRA captions use tags that were renamed upstream.
The server runs on Windows 10 or 11, on x64 or ARM64. Installation is per user and needs no administrator rights, through a setup wizard, a PowerShell one-liner or a portable zip. It can then be connected to an MCP client such as Claude Desktop, ZCode or Claude Code, either through a JSON command entry or as a plugin. Anonymous access works, so no API key is needed for the first query. A local cache serves stored answers during outages, and an upgrade command checks SHA256 hashes before replacing the binary.
Limitations and requirements
Supported only on Windows 10 or 11 (x64 or ARM64). Requests are spaced 1.5 seconds apart for anonymous access and 500 milliseconds when credentials are used.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
danbooru-tag-mcp
Danbooru MCP server for AI image generation: cached search, batch lookup, profiles, wiki, posts
v0.6.0As shown on registry.modelcontextprotocol.io when we last read it