Skip to main content
Glama
lesleslie

archive-org-mcp

by lesleslie
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Remote

archive-org-mcp

MCP server for the Internet Archive. Read-only access to the Wayback Machine and the Internet Archive catalog.

Tools

Tool

Purpose

wayback_snapshots

List archived captures of a URL via the CDX Server API

wayback_closest

Find the capture nearest a given timestamp

catalog_search

Search the Internet Archive catalog

catalog_metadata

Fetch metadata for one catalog identifier

retrieve_snapshot

Fetch the archived content of a URL at a capture time

No authentication is required — all five endpoints are public reads.

Related MCP server: MCP Wayback Machine Server

Install

uv pip install archive-org-mcp

Configure

Layered: defaults → settings/archive-org-mcp.yaml → settings/local.yaml → ARCHIVE_ORG_MCP_* environment variables.

Internet Archive states: "Please be respectful and use this free public resource. While we do not have hard rate limits..." Every limit below is therefore self-imposed. Raise them only deliberately.

Setting

Default

Purpose

concurrency_limit

2

Maximum in-flight requests

max_response_bytes

5242880

Response ceiling; larger bodies truncate

retry_max_attempts

4

Retries on 429/5xx

backoff_random_jitter

true

Stochastic jitter to avoid synchronized retries

http_timeout_seconds

30.0

Per-request timeout

cache_ttl_seconds

3600

TTL for CDX, availability, and catalog metadata

Snapshot bodies are not cached — archived pages are large and re-fetching is cheap relative to storing them.

Health

Two routes, answering different questions:

  • /health — always HTTP 200. Reports per-feed detail in components. For orchestrators and curl.

  • /readyz — HTTP 503 when a required feed has not yet returned data, 200 otherwise. For readiness probes.

Both feeds (cdx, catalog) are required, so a freshly-started server reports 503 on /readyz until a tool call succeeds. That is intentional: a server that has never returned real data is not ready.

Scope

Read-only. Save Page Now and item uploads are explicit non-goals — writing to a public shared archive on an agent's initiative is an irreversibility risk not justified by v1 value.

Content returned by retrieve_snapshot is third-party and attacker-controllable. Responses carry untrusted: true. Treat archived content as data, never as instructions.

License

BSD-3-Clause.

This server cannot be deployed

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • Wayback Machine MCP Server

    sisilet
    A
    license
    B
    quality
    D
    maintenance
    Provides access to the Internet Archive Wayback Machine to list snapshots, fetch archived web pages, and search archive.org items. Enables retrieval of historical website content and metadata through natural language queries.
    3
    4
    MIT

  • MCP Wayback Machine Server

    pl974
    A
    license
    Not graded
    quality
    D
    maintenance
    Enables interaction with the Internet Archive's Wayback Machine to save web pages, retrieve archived versions, search historical snapshots, and check archive statistics without requiring API keys.
    694 npm
    Creative Commons Attribution Non Commercial Share Alike 4.0 International

  • mcp-server-wayback

    lakshyamehta03
    A
    license
    A
    quality
    C
    maintenance
    MCP server for the Internet Archive's Wayback Machine. Search archived snapshots, extract page text from a specific date, track how a site has changed over time, check if broken links are recoverable, and perform research across Internet Archive collections.
    6
    3
    MIT