Wayback Machine — Claude Code MCP server: archives and snapshots
An MCP server and CLI for the Wayback Machine. Search snapshots, pull archived page content, compare versions — no API keys needed.
claude mcp add mcp-wayback-machine -e WAYBACK_ACCESS_KEY=<WAYBACK_ACCESS_KEY> -e WAYBACK_SECRET_KEY=<WAYBACK_SECRET_KEY> -- npx -y mcp-wayback-machineWhat Wayback Machine does
The server exposes the Wayback Machine to your agent as a set of tools. search_archives queries the CDX API, get_archived_url returns a stored copy of a page, and save_url submits a URL for archiving through SPN2. Rounding it out are list_screenshots, compare_snapshots and check_archive_status. The same operations run from a terminal: src/bin.ts figures out whether it was started as a CLI or as an MCP server and routes accordingly.
Anonymous requests work, but they share a ceiling of 15 requests per minute. Set WAYBACK_ACCESS_KEY and WAYBACK_SECRET_KEY — issued from your archive.org account — and SPN2 limits go up. Responses are cached, with TTLs chosen per endpoint: 24 hours for snapshot content, an hour for search and status, 30 minutes for save operations.
Building from source takes Node.js 22+ and pnpm. Installing through npx only needs Node 22, since the package ships compiled. One thing to know: end-to-end tests hit the live API and stay off unless WAYBACK_LIVE_TESTS is exported in your shell.
What it helps with
- see how a page looked before an edit
- list every archived version of a URL by date
- diff two snapshots of the same page
- archive a page without leaving the agent
- check whether Wayback already captured an address
Example prompts for the agent
- Find all Wayback Machine snapshots of example.com from 2023
- Fetch the archived copy of this URL and show me the text
- Save the current version of this page to the archive and tell me when it's done
How to install Wayback Machine
Platform CLI
claude mcp add mcp-wayback-machine -e WAYBACK_ACCESS_KEY=<WAYBACK_ACCESS_KEY> -e WAYBACK_SECRET_KEY=<WAYBACK_SECRET_KEY> -- npx -y mcp-wayback-machineReplace <…> with your own values
Platform CLI
codex mcp add wayback-machine --env WAYBACK_ACCESS_KEY=<WAYBACK_ACCESS_KEY> --env WAYBACK_SECRET_KEY=<WAYBACK_SECRET_KEY> -- npx -y mcp-wayback-machineReplace <…> with your own values
Config file
{
"mcpServers": {
"wayback-machine": {
"command": "npx",
"args": [
"-y",
"mcp-wayback-machine"
],
"env": {
"WAYBACK_ACCESS_KEY": "<WAYBACK_ACCESS_KEY>",
"WAYBACK_SECRET_KEY": "<WAYBACK_SECRET_KEY>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"wayback-machine": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"mcp-wayback-machine"
],
"env": {
"WAYBACK_ACCESS_KEY": "<WAYBACK_ACCESS_KEY>",
"WAYBACK_SECRET_KEY": "<WAYBACK_SECRET_KEY>"
}
}
}
}Replace <…> with your own values
Platform CLI
gemini mcp add -e WAYBACK_ACCESS_KEY=<WAYBACK_ACCESS_KEY> -e WAYBACK_SECRET_KEY=<WAYBACK_SECRET_KEY> wayback-machine npx -y mcp-wayback-machineReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Wayback Machine
How do I install Wayback Machine in Claude Code?
Run `claude mcp add mcp-wayback-machine -e WAYBACK_ACCESS_KEY=<WAYBACK_ACCESS_KEY> -e WAYBACK_SECRET_KEY=<WAYBACK_SECRET_KEY> -- npx -y mcp-wayback-machine`. The keys are optional; anonymous access works too.
What does this MCP server do?
It searches the archive for snapshots, retrieves their content, saves new page versions and compares captures. The same operations are available from the CLI.
Do I need API keys for it to work?
No, anonymous access is enough. Supplying WAYBACK_ACCESS_KEY and WAYBACK_SECRET_KEY raises SPN2 limits; the shared cap across calls is 15 requests per minute.
