Norway Archives

What Norway Archives MCP server does

Norway Archives MCP server provides read access to several Norwegian research sources through the Model Context Protocol. Digitalarkivet tools cover transcribed church records, censuses, emigrant material, and other archival records, along with the scanned pages behind those records. National Library tools search books, newspapers, journals, maps, and other digitized items. A place-name tool queries Kartverket’s official register for current names, types, and coordinates.

The server is aimed at family-history and local-history research. It can help locate a person in searchable records, identify the archive source containing a particular list, find references to a name in a digitized book, or establish the modern location of a farm or settlement.

How it works

The Digitalarkivet workflow separates discovery from evidence. da_search_persons searches only transcribed material, while da_search_sources shows whether relevant lists are scanned, searchable, or both. da_get_source describes a source and identifies where a list begins in its image sequence. da_list_pages then returns ordered pages with permanent identifiers and links. da_download_page saves a selected scan with citation information.

Record lookups include the fields in a result, the people named in it, its source, and the permanent identifier for the scanned page. National Library searches can identify pages containing a term, return snippets and printed page numbers, and provide OCR text when an item is open. Downloads create files on the user’s disk and do not overwrite existing files.

Setup and configuration

Norway Archives MCP server requires Python 3.11 or newer and uv. Run it without cloning by using uvx norway-archives-mcp, or clone the repository and start it with uv sync followed by uv run norway-archives-mcp. The process communicates over stdio, so an MCP client normally launches it.

Claude Desktop can configure the server with uvx as the command and norway-archives-mcp as its argument. Claude Code can add it with claude mcp add norway-archives -- uvx norway-archives-mcp.

No configuration is required. Optional environment variables control the cache directory, request timeout, Digitalarkivet spacing and daily budget, contact information in the User-Agent, and the download directory. A .env file is read from the directory where the server starts.

Tools and capabilities

• Search Digitalarkivet people, sources, and individual records.
• Inspect source metadata, list boundaries, page order, permanent IDs, and links.
• Download Digitalarkivet scans with archive citations.
• Search National Library collections and search within individual items.
• Read available OCR pages and download open page images.
• Search official place names and look up historical municipality and list-type codes locally.
• Inspect session requests, cache use, and the Digitalarkivet daily budget.

Limitations and notes

The Norway Archives MCP server does not maintain a family tree or write to the connected archives. Search results from Digitalarkivet are limited to transcribed lists, so an empty result does not rule out an entry in an untranscribed scan. Scanned pages should be checked before citing names, dates, or relationships.

Some National Library material is available only from Norwegian addresses. The server reports access restrictions rather than bypassing them. Digitalarkivet requests are rate-limited, cached, and subject to a daily budget; the default interval is five seconds and the default budget is 600 live requests per UTC day. National Library and Kartverket services also impose their own request limits. A bot check is reported as blocked and is not worked around.