pub search MCP#

MCP server for pub search - search ATProto publishing platforms (Leaflet, pckt, Offprint, Greengale, WhiteWind, and others using standard.site).

usage#

hosted (recommended)#

claude mcp add-json pub-search '{"type": "http", "url": "https://pub-search-by-zzstoatzz.fastmcp.app/mcp"}'

local#

run the MCP server locally with uvx:

uvx --from git+https://github.com/zzstoatzz/leaflet-search#subdirectory=mcp pub-search

to add it to claude code as a local stdio server:

claude mcp add pub-search -- uvx --from 'git+https://github.com/zzstoatzz/leaflet-search#subdirectory=mcp' pub-search

tools#

tool	description
`search`	keyword search by query, tag, platform, date, or author
`search_semantic`	semantic search by meaning, filterable by platform or author
`search_hybrid`	combined keyword + semantic search with author/platform filtering
`get_document`	retrieve full content by AT-URI
`find_similar`	find semantically similar documents
`get_tags`	list all tags with document counts
`get_popular`	see popular search queries
`get_stats`	index statistics and performance metrics

workflow#

search("space station") → [{uri: "at://...", title: "...", snippet: "...", url: "..."}]
search("gated content", author="ngerakines.me") → results from that author only
search("", author="zat.dev") → browse all docs by author
search_semantic("building a relay", author="zat.dev") → semantic search scoped to author
get_document("at://...") → {title: "...", content: "full article text..."}
find_similar("at://...") → [{uri: "at://...", title: "...", snippet: "..."}]

the author param accepts either a handle (nate.bsky.social) or a DID (did:plc:xyz). handles are resolved server-side.

development#

git clone https://github.com/zzstoatzz/leaflet-search
cd leaflet-search/mcp
uv sync
uv run pytest