Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@helmet find books about Finnish sauna and tell me which branches have them"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
HelMet library catalogue (Helsinki, Espoo, Vantaa, Kauniainen), exposed over MCP.
Same shape as tori: no cache, no DB, no scheduler.
Not a scraper — HelMet's catalogue (helmet.finna.fi) is part of Finna, run
by the National Library of Finland, which has an open REST API at
https://api.finna.fi/. No API key. HelMet records are scoped with the
building:0/Helmet/ filter. src/finna.ts is a thin client over
/search and /record.
tool | what |
| terms, page, sort, filters, facets → records with owning branches |
| Finna ID → full metadata (authors, subjects, summary, e-links, branches) |
| a title → branches that own it, formats, free-online flag, other editions |
filters are Lucene fragments taken from a previous result's facets, e.g.
{ format: "1/Book/Book/", language: "fin", search_daterange_mv: "[2015 TO 2020]", free_online_boolean: "1" }.
Related MCP server: statfin-fi
The open API returns collection holdings — which branches own a copy — not
real-time shelf status. There is no "3 available at Pasila right now" in
api.finna.fi; that lives behind the Finna web UI's authenticated ILS calls.
For live status, check helmet.finna.fi or the library.
Finna's API is meant for programmatic use. Metadata is mostly CC0/CC-BY (a few source-specific restrictions). This client only reads public catalogue data — no accounts, no personal data. House rules still apply: one request in flight, 500 ms gap, real User-Agent.
bun install
bun test # hits the live API; set OFFLINE=1 to skip
bun run mcp # stdio MCP server (src/mcp/server.ts)
bun run mcp:http # streamable-HTTP MCP server on $MCP_PORT (default 3001), for hosting
claude mcp add helmet -- bun "$(pwd)/src/mcp/server.ts" # run from the repo rootFor a remote (hosted) server: claude mcp add --transport http helmet https://<host>/mcp.
Real-time availability — needs the authenticated Finna online interface /
ILS. Would be helmet.finna.fi/AJAX/JSON?method=getItemStatuses scraping;
skipped on purpose.
Personal account (loans, holds, fees, renew) — needs per-user credentials. Different tool, different legal posture. Not wired up.
No cache / scheduler / DB — stateless by design; add if a request ever needs it.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
No tool schema history has been recorded yet.
This server cannot be installed
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Search books and authors, fetch editions, browse subjects, and resolve cover images.
Search and fetch Wikidata entities, execute SPARQL queries, and resolve external identifiers.
Search and query government open-data portals (Socrata SODA API).
Federated search of books and papers, BibTeX/RIS citations, open-access retrieval and reading.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/zeriaxdev/helmet-scraper'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
