You hit a paywall on the one paper you need. Elsewhere asks the public indexes whether an open copy of that same article already exists: an archive capture, a wire reprint, a repository deposit. It reports a hit only when it has confirmed a readable copy.
free tier: 5 searches a day · pro: 3 browser seats · no subscription
journalofaridstudies.example · subscription required · DOI 10.1234/jas.2021.0417
Fetched and checked: full article text, 9,140 words, PDF.
repository.example.ac.uk/items/8841/rainfall-sahel.pdf2 unverified links kept · nothing was bypassed
Elsewhere reads the page you are already on for its identifiers, then goes looking somewhere else entirely. The article you cannot open stays exactly as unopened as it was.
Elsewhere spots the usual paywall furniture: a truncated article body, a subscribe interstitial, a metered-access banner. On Pro you can have it start searching the moment it sees one. On Free you click the icon.
The URL, the headline, the publication date, the DOI when the page carries one. That is the search key. No credentials, no cookies, nothing from behind the wall.
Up to five routes in parallel, each one a public index of material that was already published openly. Free gets two of them. Pro gets all five.
Each candidate is fetched and inspected for readable article text. Confirmed copies are reported as found. Everything else is labelled unverified, with the link kept so you can look yourself.
Every route has a failure mode. Here they are, written down, so the extension never has to surprise you.
The Internet Archive holds captures of an enormous slice of the web, going back decades. For news and for pages that were once open and later closed, it is the single most productive route Elsewhere has.
Wire copy from AP, Reuters and PA is reprinted, often word for word, by outlets that do not charge. If the piece you want started life on a wire, a free version of it is usually sitting on three other sites.
For anything with a DOI, Unpaywall resolves to legally deposited open-access versions: author manuscripts in university repositories, publisher open-access copies, PubMed Central. This is the strongest route for academic reading and the one researchers come for.
Human-submitted snapshots, which makes the coverage of paywalled news reporting better than any crawler manages. When somebody archived the story you want, they usually archived a readable version of it.
Some publishers still serve an AMP version of their pages that was never wired into the metering. Where that is true it is instant and clean, and it costs one request to check.
Most tools in this space report a result the moment a URL comes back from an index, which is how you end up clicking through to another subscribe box. Elsewhere fetches the candidate and looks at what arrived. Three verdicts, and it never blurs them.
Elsewhere opened the copy and readable article text came back. This is the only verdict that counts towards the number in the badge.
A copy very likely exists, but the check could not complete: a fetch was blocked, a redirect went somewhere odd, the text was too short to trust. You get the link and the reason, and you decide.
The route came back empty. Elsewhere says so plainly rather than padding the result list, and you have lost about two seconds finding out.
The reason this matters: a tool that overstates its results is a tool you stop trusting after the third dead link. Elsewhere would rather report one confirmed copy and two maybes than five results you have to check yourself.
Your library subscription does not cover every journal on earth, and inter-library loan takes days. Unpaywall finds the deposited manuscript in minutes, and it is the version the authors were allowed to post.
Unpaywall route, ProA reading list of forty items, of which six are behind walls your institution does not pay for. Check all six in the time it takes to email a lecturer about one.
All five routes, ProYou need the story you are citing, and you need it now. Wire pieces are reprinted free somewhere almost every time, and the syndication search is on the free tier.
Syndication and Wayback, FreeFive searches a day, free, forever. That covers most people's browsing. Pro is for the days when five is not enough.
Free tierNothing Elsewhere does is secret. Every route is a public service you can visit in a browser tab. The difference is how long five tabs take, and whether you actually open the fifth.
| Step | By hand | With Elsewhere |
|---|---|---|
| Find the DOI on the page | Scroll, squint, copy | Read from the page metadata |
| Check five routes | Five tabs, five queries | One click, run in parallel |
| Confirm a copy is readable | Open each result and read it | Fetched and checked before you see it |
| Know a route found nothing | Only if you got that far | Stated for every route, every time |
| Typical time | Three to five minutes | Two to four seconds |
Timings are our own measurements on a normal connection, not a benchmark. The full comparison, including how Elsewhere differs from paywall-bypass extensions.
The free tier is the real extension with a daily cap and two routes. Pro removes the cap, adds the three routes that need more work behind them, and covers three browsers for life.
No key, no account, no time limit
Everything in Free, uncapped
Chrome, Edge and Brave. Your searches stay on your machine; the only thing the extension sends us is your licence key at activation.
Send your details and we will email the download, your key and the invoice, usually within the hour. One key activates on three browsers.
Questions first? Email simon@valleypc.co.uk and a person will answer.
No, and it is not built to. Elsewhere never touches the paywalled page. It takes the article URL, title and DOI, then asks public indexes whether an openly published copy of that same work exists somewhere else: a web archive capture, a syndicated wire reprint, a university repository deposit. If one exists, you get the link. If none exists, Elsewhere tells you so and you are back where you started.
A hit means Elsewhere fetched the candidate copy and confirmed readable article text came back. Anything it could not confirm is labelled unverified, and you still get the link so you can judge it yourself. The count in the badge only ever includes confirmed copies.
Unpaywall's public API requires a contact email on every request. That is their rule, not ours. You type it once in the extension settings, it is sent only to Unpaywall as part of the query, and Elsewhere stores it nowhere else.
archive.today sits behind Cloudflare, which blocks automated fetches. Elsewhere can tell you a snapshot address exists and hand you the link, but it usually cannot open the page to confirm the text. Marking that as a confirmed hit would be a lie, so it does not.
It shrinks every year. Google deprecated AMP as a ranking signal and publishers have been removing the markup ever since. The route still helps on the outlets that never got round to it, and it is off the road the day it stops paying its way.
Five searches a day across the Wayback Machine and syndication search. Same verdict logic, same honesty about what was confirmed. Pro removes the daily cap, adds Unpaywall, archive.today and AMP, and turns on auto-search when a paywall is detected.
One licence key activates Elsewhere Pro on up to three browser installs, for example Chrome on your laptop, Chrome on your desktop and Edge. Switching machines? Release the old seat yourself on the manage page.
No. $39 once, including future updates. If it is not for you, there is a 14-day money-back guarantee.