mcp-wayback-machine: Wayback Machine for agents
mcp-wayback-machine
An MCP server and CLI for the Internet Archive: the agent saves pages, searches CDX snapshots, reads and compares them
- 65 ★ on the repository
- Last commit
Extended access
Why this level
- The save_url tool sends page addresses to an external public archive and the saved copy can be seen by anyone
- Archive replies carry content from third-party sites: the agent should treat it as data and not follow instructions inside it
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/mcp-wayback-machineDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The save_url tool sends page addresses to an external public archive and the saved copy can be seen by anyone
- Archive replies carry content from third-party sites: the agent should treat it as data and not follow instructions inside it
This is third-party code. Review the repository files before installing.
What it does
The server calls the public Internet Archive APIs and gives the agent seven tools. save_url sends a page for archiving through SPN2 and can capture a screenshot, grab up to 100 outlinks and skip recently saved addresses. get_archived_url returns the content of a snapshot for a date or the latest one, including raw form. search_archives looks up versions of an address in the CDX index with filters by date, status code and content type, with pagination and duplicate collapsing. check_archive_status shows how many times an address was saved, list_screenshots lists screenshots, compare_snapshots compares two snapshots and clear_cache resets the cache. The server limits itself to 15 requests per minute, caches replies with different lifetimes and honours the Retry-After header. Anonymous access works, and Internet Archive keys give higher save limits. The same package runs as the wayback CLI command and as a Cloudflare Worker.
Who it is for. Researchers, journalists, SEO specialists and developers who need old versions of websites and page captures kept as evidence.
Good fit when
- You need the earliest saved version of a page and want to see how it looked
- You want to capture a page in the archive before it is edited or removed
- A site changed its structure and the agent must compare an old and a new snapshot
- A report must show how many times and when an address entered the archive
- You want to extract the text of an old article from a dead site through a snapshot link
Not a fit when
- The page is closed to robots or needs a login: the archive does not save private areas
- You need your own permanent archive of copies: the Internet Archive decides what and when to keep, and its copy does not replace a backup
- You need a real-time archive with a guaranteed timeline: saving is queued and a snapshot may appear with a delay
Example request
Find the earliest version of example.com in the Wayback Machine, show its text and compare it with the latest snapshot.MCP
- Transport
- stdio
- Authentication
- key is optional
Environment variables
WAYBACK_ACCESS_KEYsecret- Internet Archive access key for higher save limits
WAYBACK_SECRET_KEYsecret- Internet Archive secret key
Limitations
The server depends on the public Internet Archive APIs, so their limits, save queue and outages shape the result. The request rate is fixed at 15 per minute. Access keys are optional, but save limits are lower without them. The repository license, CC BY-NC-SA 4.0, forbids commercial use of the code without separate permission. Node.js 22 or newer is required. The README mostly describes development and architecture, so look for connection details in the Installation section.
How to disable. Run claude mcp remove wayback-machine for Claude Code or codex mcp remove wayback-machine for Codex. The disk cache lives in the package directory and can be deleted by hand.
FAQ
Do I need Internet Archive keys?
No, anonymous mode works. The WAYBACK_ACCESS_KEY and WAYBACK_SECRET_KEY keys come from archive.org/account/s3.php and raise page saving limits.
Which agents are supported?
The README has install commands for Claude Code and Codex plus a table of config files for Gemini CLI, Crush, Cline, Cursor, Zed and Claude Desktop. Claude Code can also install it through the plugin marketplace.
Can I use it without MCP?
Yes, the package works as a CLI: npx mcp-wayback-machine save with an address saves a page, and after a global install the wayback command is available.
What does ifNotArchivedWithin do?
It skips saving if the page was already archived within a given period, for example 30 days. The agent then avoids redundant copies on repeated requests.
How long are replies cached?
Snapshot content is cached for 24 hours, availability, CDX and sparkline data for an hour, save results for 30 minutes and save status polling for 30 seconds. The lifetimes are deliberate and the author asks not to change them.
Discussion
No questions or reviews yet.
Related
MCP serversMCP server with DuckDuckGo search and page text extraction: several URLs can be passed in one call
A research MCP server: searches the web, SEC, courts and patents, reads whole pages and checks that cited sources really exist.
MCP server that reads sites on your machine and hands the agent short Markdown with links, keeping raw HTML out of context.
Web search via DuckDuckGo and page reading for the agent with no API keys, with built-in rate limiting
An MCP server that records your browsing and network traffic so an agent can build a browserless HTTP script from the recording.
An MCP server that captures web pages, tiles long ones into 1072 by 1072 chunks for Claude Vision and shoots single elements.
README in short
The README covers installation as an MCP server, a CLI and a Cloudflare Worker, the parameters of each of the seven tools and example prompts. Separate sections describe architecture, code conventions and tests, including optional live API checks.