Site History
A time machine for your own website
Go back to any day and see, open and measure the page as it was. We visit your pages on a schedule, keep everything, and make the history available to AI agents through a skill, an MCP server and a CLI.
What one capture stores
Every artifact comes from one page load, so the HTML, the rendered page and the screenshot describe the same visit. A capture is never changed after it is written.
| Artifact | What it is |
|---|---|
manifest.json | IDs, UTC times, profile, status and a SHA-256 for every artifact, so each capture can be cited and verified |
raw.html | The main document exactly as delivered, what a non-rendering crawler gets |
rendered.html | The DOM after the page settled, what a browser shows |
archive | Every request and response of the page load, a replayable working copy |
screenshot-full and screenshot-fold | Full page (lossy WebP, in tiles for tall pages) and above the fold (lossless WebP) |
response.json, network.json, console.json, metrics.json | Redirects and headers, request log, console errors, totals |
page.md | Main content as clean Markdown, cheap for a model to read |
Profiles: desktop and phone by default, and optionally a Googlebot user agent. Lighthouse and Chrome UX Report data of the day are available with the Performance add-on. Plugins turn the history into named change events such as a relaunch, moved URLs or a broken page.
Who it is for
SEO and analytics consultancies
Page history of client sites for AI analysis, and evidence for audits.
Large sites with many releases
Know what changed and when, catch SEO relevant regressions the day they ship, settle "we did not change anything".
Mid-size sites and shops
A daily record of what the important pages delivered, with alerts when title, canonical, indexability or payload change.
Agencies
One account, many client projects.
AI first: MCP, CLI and skill
One API, with the MCP server, the CLI and the Claude skill as thin clients. Answers are small and structured, with pointers to the files, so an agent asks for the timeline first and opens only what it needs.
MCP server
Tools such as as_of, timeline, diff and verify_claim.
CLI
snap timeline, snap get, snap pull for scripts and agents without MCP.
Skill
Teaches an agent when to use the service and to cite capture ID, UTC time, profile and hash for every claim.
Honest limits
- A capture is one defined visitor, observed over a recorded interval. A first-time visitor from Germany with a given profile. It is not proof of what every user saw, and nothing is known about the time between two captures. Missed days show as gaps.
- The Googlebot profiles are not Googlebot. They send a Googlebot user agent from our IP address. They show what your site sends to such a request, but not what real Googlebot gets where your site verifies Googlebot by IP or reverse DNS.
- Your own domains only. We capture only domains you have verified. See how verification works.
- Protection may block us. A blocked capture is detected and reported, never stored as if it were your page. See allowlisting.
- Prices are a proposal that is still to be tested with real buyers.
See what your page looked like last month
Start with the free plan: one URL, captured daily, kept for 16 months. A verified domain is required.
Sign up free