dnb-mcp 📚

mcp-name: io.github.mohammad-abbasi-sh/dnb-mcp

Unofficial read-only MCP server for services.dnb.de — Deutsche Nationalbibliothek catalogue via its official SRU API: keyword, title and author search, ISBN and DNB-ID lookup. No key, no login. No key, no login.

Built by Mohammad Abbasi as part of the "Germany MCP factory". Not affiliated with services.dnb.de.

For AI agents

Add to your MCP client config (stdio):

{
  "mcpServers": {
    "dnb": {
      "command": "uvx",
      "args": ["dnb-mcp==0.1.0"]
    }
  }
}

Or run locally: python -m dnb_mcp.server (needs mcp<2, httpx).

Tools

Tool	Description
dnb_search_books	Keyword search across the DNB catalogue — query, limit 1..50
dnb_search_by_title	Title search — title, limit 1..50
dnb_search_by_author	Author search — author (e.g. "Kafka, Franz"), limit 1..50
dnb_by_isbn	Look up one book by ISBN — isbn (13 or 10 digits, with or without dashes)
dnb_get_record	Full record by DNB identifier — dnb_id (e.g. "1306459214")
dnb_health	Reachability probe (ok / blocked / error) — one polite request

Notes

• All data comes from the Deutsche Nationalbibliothek SRU API (free, no key, no login).
• Indexes used (from the API's own explain response): woe (all terms), tit (title), per (person/author), num (number incl. ISBN), idn (DNB identifier).
• There is no newspapers-only filter: the mat (material type) index returns 0 for every German value tried (Bücher, Zeitschriften, Zeitung, Zeitungen, Online-Ressource), so no newspapers-only tool is offered; newspapers are still found by the keyword and title searches.
• Dublin Core recordSchema is unsupported by the API — MARC21-xml only.

Tests: 20/20 green (14 offline + 6 live, 2026-11-10).

Development

python3 -m venv .venv && .venv/bin/pip install "mcp<2" httpx pytest
.venv/bin/pip install -e .
.venv/bin/python -m pytest tests/ -q

License

MIT

---

Built by Mohammad Abbasi — Almanibama · mabbasi.ir