Archivepanel Web Preservation System

Web Preservation System

Links rot.
Archives don't.

Archivepanel captures a page into a single self-contained HTML file — every stylesheet, image and font embedded, every script stripped — and files it in your private account. Nothing is ever rewritten, and nothing depends on the original site still being there.

Free tier, no card required/ 50 MB of snapshot storage/ Two-factor sign-in on every account

Per capture
One file
The whole page in a single document. No folder of assets, no broken references.
Stored artifacts
Five
Document, preview image, plain-text rendering, WARC container and a JSON manifest.
Scripts retained
Zero
Stripped at capture, and a policy embedded so an opened archive cannot reach the network.
Interchange
ISO 28500
WARC 1.1 export that opens in pywb and ReplayWeb.page.

The problem

A bookmark is a promise someone else has to keep.

Pages get edited, paywalled, redirected and deleted. Screenshots lose the text. "Print to PDF" loses the layout. A saved page dumps a folder of assets that half-loads a year later. And a link in a citation, a contract, a compliance file or a dispute is only worth what the other side's server still says today.

Evidence

What the page said, on the day

Terms, prices, policy statements, published claims. Each snapshot is stamped with the address it came from and the moment it was taken, and it is never modified afterwards.

Research

Sources that survive the citation

Capture the page you cited, keep the plain text beside it for quoting and grepping, and export WARC when an institution accepts nothing else.

Monitoring

A record that keeps itself

Point a schedule at a page — or at a sitemap, and get the whole section — and the history builds without anyone remembering to press a button.

How a capture works

Five steps, no dependency on the site afterwards.

01

Screen

The address is validated first. Anything resolving to loopback, private, link-local or carrier-grade NAT space is refused outright.

02

Capture

Headless Chromium loads the page, dismisses a cookie banner if it recognises one, and waits for the page to settle. Stylesheets, images and fonts are embedded, unused CSS pruned, scripts stripped.

03

Verify

The preview is a screenshot of the saved file, not the live page — so a capture that lost its styling looks wrong immediately, not a year later.

04

Render

A structured plain-text version is derived from the same bytes, so the text and the archive can never disagree about what the page said.

05

File

Document, preview, text, WARC and manifest are written to private storage under your account, and counted against your allowance.

Archives are immutable. A capture records what a page said on a given date, so nothing ever rewrites one. "Capture again" files a new snapshot alongside the old. Deleting is the only way to lose one, and it takes your password. A capture still queued or running can be stopped, and a stopped one writes nothing at all — so abandoning it costs you no storage.

What you get

Every capture leaves five files behind.

Not one format you have to live with. The document for reading, the image for recognising, the text for quoting, the container for everyone else's tools, and a manifest that accounts for all four.

FileWhat it isWhat it is for
.html Self-contained document The page itself, assets inlined as data URIs and scripting removed. Opens in any browser with no network access at all.
.jpg Preview image Rendered from the saved file. Makes a snapshot recognisable in a list, and proves the capture actually kept its styling.
.txt Plain-text snapshot Headings underlined, lists bulleted, tables drawn as an ASCII grid, link targets printed beside their text. Read, quote, grep and diff without opening a megabyte of HTML.
.warc.gz WARC 1.1 container The interchange format the rest of the web-archiving world reads. Downloads as one file carrying the document, the text and the screenshot together.
.json Capture manifest A size, a SHA-256 digest and a content type for every stored file, plus the addresses, times, viewport and tool versions behind the capture. An inventory or an integrity check without opening the snapshot.

Capabilities

Built like an operations tool, not a bookmark folder.

Scheduled captures

Hourly, daily, weekly or monthly

A page schedule captures its own address. A sitemap schedule reads the address as an XML sitemap and captures the pages it lists — index files included — so one schedule covers a whole section of a site and keeps covering it as the site grows.

Live progress

You do not have to wait on the page

Status, stage and progress update in place while a capture runs. When one finishes, fails or is stopped by a quota, a notification is waiting in the tray — and arrives as a toast if you are still looking.

Text snapshots

Readable, quotable, diffable

Every capture is also rendered to structured plain text from the stored bytes. Compare one week to the next, quote a paragraph, or search a snapshot without loading it.

WARC export

No lock-in, by design

Every capture is packaged as a WARC file so a snapshot taken here is not trapped here. It replays in ReplayWeb.page and pywb, and satisfies an archive that accepts nothing else.

Private storage

Readable only through your account

The bucket is private and no signed URL is ever handed out. Viewing, downloading and WARC export each stream the object from behind the session gate, so a link to an archive is worthless to anyone else.

Activity log

An account that can be audited

Sign-ins, captures, schedule changes and deletions are recorded, so the question "who did what to this record, and when" has an answer.

Sites

The history of one place, not one page

Every snapshot you hold is grouped by host. Searching inside a site reads the stored text of its snapshots and shows you the passage that matched, so a page can be found by something it said rather than by its title.

Consent banners

The page, not the overlay

A cookie banner covers the page and usually locks scrolling with it, so everything below the fold never loads. Where a known consent platform is recognised, its own button is pressed before the page is saved — and the record says what was done about it.

Provenance

The moment, not just the page

Each record carries what the page declared about itself — language, canonical address, author, published date — and how its host resolved at that moment: the addresses it answered on, the CNAME behind a CDN, the authoritative nameservers and the certificate it served.

Security

Captured markup is hostile input. It is treated that way.

An archive is somebody else's HTML, stored on our infrastructure and later opened in your browser. Most of the engineering here is about making that safe rather than merely convenient.

The capturer cannot be pointed inward. Every address — typed, scheduled, or named by a sitemap — is screened before it is fetched, and again at each redirect hop.

Account and archive protection

  • Two-step sign-in. Password, then a six-digit code sent to your address. Only the code's digest is stored.
  • argon2id password hashing at 64 MB with a unique salt per hash.
  • Session tokens are stored as digests, so a database disclosure yields no usable sessions.
  • Archives render sandboxed. A stored document is served under a policy that pins it to an opaque origin with scripts and forms refused.
  • Every query is scoped to your account. There is no cross-account read path.
  • Deleting requires your password and is rate limited — it is the only irreversible action in the product.

Capture your first page in about a minute.

Create an account on the free tier, paste an address, and watch the snapshot land. Every archive is exportable as WARC from the moment it exists.