Skip to main content
Glama
archival-imagery-mcp

archival-imagery-mcp

by chandhoke

archival-imagery-mcp

An MCP server for browsing and downloading open-licensed imagery from six major museum and archive APIs. Built for editorial projects that need authored, archival visual material instead of AI-generated stock.

License: MIT Node 18+ MCP

What this is

When Claude or any other MCP-compatible client needs imagery for an editorial project — a service site, a long-form article, a brand book, a deck — generating images from scratch usually produces something that reads as AI. For brands with editorial bona fides, that signal is fatal.

This server gives the assistant a different option: search the world's open-access museum collections for authored, attributed, often-centuries-old imagery and download it straight into the project's public/ folder. Most results are CC0, CC-BY, or Public Domain Mark — usable with no licensing friction.

Related MCP server: IIIF MCP Server

Sources

Source

Coverage

API Key

License Defaults

Wellcome Collection

UK. Medical, scientific, esoteric, astrological history. ~250k items.

No

CC-BY / PDM

Met Museum

US. Open Access program. ~470k objects, all CC0 when public domain.

No

CC0

Library of Congress

US. Photos, manuscripts, maps, prints, newspapers, film. ~1M items.

No

Mostly PD

Smithsonian Open Access

US. 4.5M+ CC0 items across 21 museums.

Free

CC0

Europeana

EU aggregator — British Library, Louvre, full Rijksmuseum collection, plus 4000+ institutions.

Free

Mixed (filterable)

Note on Rijksmuseum: the Rijksmuseum's REST API was deprecated in 2024 in favour of OAI-PMH (an XML harvesting protocol designed for institutional bulk-download, not query-by-keyword). Rather than re-implement on a less query-friendly protocol, this server intentionally omits direct Rijksmuseum support. Europeana fully indexes the Rijksmuseum collection — query it via europeana_search with provider="Rijksmuseum".

Installation

No install step needed if you have Node 18+ — just point your MCP client at npx:

{
  "mcpServers": {
    "archival-imagery": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "archival-imagery-mcp"],
      "env": {
        "SMITHSONIAN_API_KEY": "your-key-here-or-omit",
        "EUROPEANA_API_KEY": "your-key-here-or-omit"
      }
    }
  }
}

Or install once globally:

npm install -g archival-imagery-mcp

Then use "command": "archival-imagery-mcp" (no args needed) in your config.

API keys are all optional — Wellcome, Met, and Library of Congress work without keys. Smithsonian and Europeana tools return a helpful error message with the signup URL if their key is missing.

From source (for development / custom builds)

git clone https://github.com/chandhoke/archival-imagery-mcp.git
cd archival-imagery-mcp
npm install

Then point your MCP config at the local path:

{
  "mcpServers": {
    "archival-imagery": {
      "type": "stdio",
      "command": "node",
      "args": ["/absolute/path/to/archival-imagery-mcp/index.mjs"]
    }
  }
}

Getting the optional API keys

All free, all take under 2 minutes:

  • Smithsonian: api.data.gov/signup/ — email confirmation. The same key works for NASA, NOAA, FCC, and most US government APIs.

  • Europeana: pro.europeana.eu/page/get-api — instant. Covers British Library, Rijksmuseum, Louvre, and 4000+ other institutions in one endpoint.

Tools (12 total)

Search & retrieval

Tool

Source

Needs Key

wellcome_search

Wellcome

—

wellcome_get_work

Wellcome

—

wellcome_image_url

Wellcome (utility)

—

met_search

Met

—

met_get_object

Met

—

loc_search

Library of Congress

—

loc_get_item

Library of Congress

—

smithsonian_search

Smithsonian

✓

smithsonian_get_object

Smithsonian

✓

europeana_search

Europeana

✓

europeana_get_record

Europeana

✓

Shared utility

Tool

Purpose

download_image

Save any image URL (from any of the above) to a local path. Creates parent dirs automatically.

Usage examples (natural language)

Once installed, just describe what you want and the assistant picks the right tool:

"Find Indian astrology pieces at the Met that we could use on /method — show me 5."

"Search Wellcome for nakshatra diagrams and save the best two to my project's public/brand/ folder."

"Use Europeana with provider='Rijksmuseum' to find Vermeer paintings — I want a portrait for the about page."

"Use Europeana to find anything from the British Library tagged 'Sanskrit manuscript' — only CC0/CC-BY please."

Attribution

When using imagery from any of these sources in published work, always check the per-item license (returned with every result) and follow its attribution requirements. Most CC-BY items only need a short credit line:

Imagery courtesy of [Source] (CC BY 4.0).

CC0 / Public Domain Mark items have no attribution requirement but crediting the source is good practice.

Architecture

  • Single-file MCP server (index.mjs), ~400 lines

  • Native fetch (Node 18+), no HTTP library dependency

  • Only one dependency: @modelcontextprotocol/sdk

  • No telemetry, no persistent state, no auth required for the server itself

  • Stateless — every tool call hits the upstream API fresh

Extending

To add another source (e.g. New York Public Library Digital Collections, DPLA, Gallica):

  1. Add a section under // ========================================================================= with xyzSearch, xyzGetObject functions

  2. Add an env-var helper if it needs a key (use the existing requireKey pattern)

  3. Add entries to the TOOLS array with clear descriptions

  4. Add entries to the HANDLERS map

~50 lines per new source.

License

MIT © Harpreet Chandhoke

This server is a client for third-party APIs. Each source has its own terms — respect them in your usage. The MIT license applies only to this server code, not to the content retrieved.

Available Tools

12 tools
download_imageA

Download any image URL to a local file path. Parent directories are created automatically. Works with images returned from any of the above sources.

Inputs: url (required), savePath (required).

europeana_get_recordA

Fetch a Europeana record by ID (e.g. '/2048128/618580'). Requires EUROPEANA_API_KEY.

Inputs: recordId (required).

loc_get_itemA

Fetch a Library of Congress item by full URL (e.g. 'https://www.loc.gov/item/2003663891/'). Returns metadata and all image file URLs.

Inputs: itemUrl (required).

met_get_objectA

Fetch a Met object by numeric ID with all image URLs (primary + additionals) and metadata.

Inputs: objectId (required).

smithsonian_get_objectB

Fetch a Smithsonian object by ID with all image URLs and metadata.

Inputs: id (required).

wellcome_get_workA

Fetch a Wellcome work by ID with all associated IIIF image URLs.

Inputs: id (required).

wellcome_image_urlA

Build a direct Wellcome IIIF image URL at any IIIF Image API size ('1200,', ',900', 'full', etc.).

Inputs: iiifId (required), size.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4/5.0

    Scored across 12 tools

    Disambiguation5/5

    Each tool is clearly scoped to a specific institution and action (search vs. fetch vs. utility). The institution prefixes make it immediately clear which API a tool targets, and search/fetch pairs within each institution are unambiguous. Even the generic download_image is distinct as a utility.

    Naming Consistency4/5

    The naming follows a predictable {institution}_{search|get_*} pattern for the source-specific tools, with snake_case throughout. The only deviation is wellcome_image_url, which uses a noun phrase instead of a verb, but it is still recognizable and consistent in style.

    Tool Count5/5

    12 tools is well-scoped for a multi-institution archival imagery server. Each institution has at least a search and a fetch tool, plus a generic download utility and a IIIF URL builder. No tool feels redundant or unnecessary.

    Completeness5/5

    The server covers the complete lifecycle for the domain: search each institution's collections, retrieve detailed metadata and image URLs, and download images. All included institutions have both search and fetch capabilities, so there are no dead ends or obvious missing operations.

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      B
      quality
      D
      maintenance
      An MCP server that enables searching and fetching openly-licensed images from Openverse with features like filtering by license type, getting image details, and finding essay-specific illustrations.
      5
      157 npm
      20
      MIT
    • A
      license
      A
      quality
      D
      maintenance
      MCP server for searching museum collections and viewing artwork images and metadata from multiple museums, including the Met, Art Institute of Chicago, Rijksmuseum, and more.
      4
      11 npm
      2
      MIT