The copyright notice on a website tells you who owns the material and the year protection started — that's it. It does not grant you permission to copy, republish, scrape, or reuse anything on the site. What you're actually allowed to do depends on four separate things: the notice itself, the site's terms of service, any license attached to specific content (Creative Commons, public domain marks), and the statutory exceptions in your country, such as US fair use under Section 107 of the Copyright Act. A "© 2026" footer with no license language means all rights are reserved by default.
Most readers think the copyright line is the permission. It isn't. The permission lives somewhere else, and if you don't find it, the default under the Berne Convention — which the US joined in 1989 and which covers 181 contracting parties as of the World Intellectual Property Organization's count — is that no reuse is authorized. Let's walk through what you actually have a right to do, section by section, with the specific limits that matter.
Read the Three Separate Documents, Because They Do Different Jobs
Perplexity, Harvard's Office of the General Counsel, and most university library guides all say the same thing: the copyright notice, the terms of service, and any content license are three different instruments. The notice identifies ownership. The terms of service govern your conduct on the site — what you can do with an account, how you can access the content, whether scraping is permitted. A content license (often Creative Commons) grants affirmative reuse rights over specific works.
Here's the failure case that catches people: a site can display a Creative Commons license on its blog posts but reserve all rights on its photography. Scraping the photos because the blog was CC-licensed gets you a takedown notice under the Digital Millennium Copyright Act, not a polite email. The license attaches to individual works, not to the domain.
A second trap is the terms of service operating as a contract independent of copyright. Even where fair use would let you copy something, the site's terms might prohibit it — and breach of contract is a separate claim from infringement. Courts have split on whether browsewrap terms (terms you never click to accept) are enforceable; clickwrap terms, where you affirmatively agree at signup, generally are. So check how you agreed.
Third, archive and reprint sites often carry a "reprint rights" page that's buried under a "Permissions" footer link rather than next to the copyright symbol. If you're planning commercial reuse, the permissions page is the first place to look, before you ever read the copyright line. If you can't find one, assume the answer is no.
A Copyright Notice Confers No Rights — It Only Records Ownership
Under US law, works published on or after March 1, 1989, no longer require a copyright notice to be protected. So a page with no © symbol is still fully copyrighted. The notice's practical value is evidentiary: 17 U.S.C. § 401(d) says that where a proper notice appears, a defendant in an infringement suit cannot claim "innocent infringement" to reduce statutory damages. That's why statutory damages can run from $750 to $30,000 per work for ordinary infringement and up to $150,000 for willful infringement, per 17 U.S.C. § 504(c).
That $750–$150,000 range is the number most footer-reading guides skip. It's the actual financial exposure. If you republish a single photograph without permission and the photographer registered it with the US Copyright Office — registration matters, because § 412 bars statutory damages and attorney's fees for unregistered works — you're looking at a range starting at $750 and climbing based on willfulness. The Copyright Alliance reports that statutory damages are the primary leverage in these disputes.
The notice format itself isn't a magic formula, but the standard is the word "Copyright" or the © symbol, the year of first publication, and the name of the owner. Some sites also list a range ("© 2018–2026"), which just means the site has been running since 2018 with updates. It doesn't mean the older content is expiring. Under current US law, copyright for works created on or after January 1, 1978, lasts the life of the author plus 70 years.
Absent a License, the Default Is "All Rights Reserved" — Including for Print and Screenshots
If a site has no license language at all, the default is that the copyright owner reserves every exclusive right under 17 U.S.C. § 106: reproduction, derivative works, distribution, public performance, and public display. That's five separate rights, and each one needs its own permission or comes under its own exception. Printing a page for personal use is legally low-risk in practice — it's a single reproduction — but that's a practice norm, not a legal grant.
The screenshot question is the one readers most often get wrong. Screenshotting a site for a critique, a tutorial, a review, or news commentary usually sits closer to fair use than a screenshot used as decorative imagery on a marketing page. The determining factors are the four in § 107: purpose and character of the use (commercial vs. educational, transformative vs. substitutive), nature of the work (factual vs. creative), amount used (a whole page vs. a crop), and effect on the market for the original. A full-page screenshot used to illustrate a review of that site is defensible. The same screenshot on a landing page trying to look authoritative is not.
Screen-scraping at scale is a different category entirely. The Ninth Circuit's hiQ Labs v. LinkedIn decisions (2019 and 2022) addressed scraping of public data under the Computer Fraud and Abuse Act, not copyright — and hiQ ultimately settled and the case was vacated. So don't cite hiQ as permission to scrape. Even where scraping doesn't violate the CFAA, the copied content is still subject to copyright, and the terms of service usually prohibit it in writing.
| Action | Default legal status (no license) | What usually triggers a claim |
|---|---|---|
| Personal reading and printing | Low practical risk | Redistribution or commercial sale |
| Full-page screenshot in a review | Fair use analysis required | Use as decorative imagery elsewhere |
| Republishing an article excerpt | Infringement absent fair use | Republication beyond a short quotation |
| Scraping public pages | Governed by ToS + copyright | Bulk copying into a competing product |
| Using images from the site | Infringement | Almost always, absent permission |
| Linking to the page | Usually fine | Framing the page to hide its source |
The last row matters. Hyperlinking is generally not copyright infringement because it doesn't copy the work. Framing — embedding the page inside your own layout so the URL doesn't change — has been litigated and outcomes depend on jurisdiction; the safest version is a plain link.
Fair Use Has Four Factors and a Number You Probably Haven't Heard
US fair use lives in 17 U.S.C. § 107 and has no bright-line word count, percentage, or "x seconds of audio" rule. The four factors are purpose and character, nature of the work, amount used, and market effect. The commonly cited threshold that circulates in publishing — that a work has entered the "public domain" if it was published before 1929 as of 2025, moving to 1930 as of January 1, 2026 — applies only to US publication dates under the Copyright Term Extension Act. It has nothing to do with how much of a current site you can quote.
The factor most people ignore is the fourth: market effect. A short quotation in a review generally helps the market for the original. A short quotation that gives away the article's core value — the one answer the reader needed — replaces the market for the original even if it's ten words long. Length isn't the metric; substitution is.
Fair use also varies by jurisdiction. The UK has "fair dealing" for specific purposes (criticism, review, news reporting, research, parody) with no open-ended fair use doctrine; Canada's Copyright Act has a fair dealing provision interpreted more broadly after CCH Canadian Ltd. v. Law Society of Upper Canada (2004); the EU's Directive 2019/790 introduced a specific text-and-data-mining exception. If you're quoting a US site from outside the US, the applicable law is usually yours for your conduct, not the site's — but the site can still enforce its rights where it's registered.
Creative Commons and Public Domain Marks Override the Default, With Conditions
A Creative Commons license is the most common affirmative grant you'll see on a site. The licenses are versioned — CC BY 4.0 is the current international version — and every version imposes different conditions. The four license elements are attribution (BY), share-alike (SA), non-commercial (NC), and no-derivatives (ND). Two combinations to remember: CC BY-NC-ND 4.0 means you can redistribute unmodified copies with attribution, and that's all; CC BY-SA 4.0 means adaptations must carry the same license.
Attribution under CC BY 4.0 has specific requirements in Section 3(a): you must identify the creator, provide the license name and a link to it, indicate whether changes were made, and retain any copyright notice. "Photo: [site]" is not compliant attribution. You need the creator's name, the license identifier, and a link.
Public domain marks are not the same as CC0. CC0 is a waiver — a legal tool by which the creator gives up rights to the extent the law allows. The public domain is a status of a work whose copyright has expired or was never valid. The distinction matters when you're building a database and need to prove provenance: CC0 has a waiver trail; public domain relies on date arithmetic. Wikimedia Commons and the Internet Archive both maintain separate tags for the two.
Register the Failure Case: A Takedown That Ruins a Monday
Walk through a realistic scenario. You run a small newsletter and you lift an infographic from a site to illustrate a story. The site has no visible license, no © symbol, and a standard footer. Under US law, that footer means the site reserves all rights. The infographic's creator registered it three years ago with the US Copyright Office. Six weeks later, you receive a DMCA notice under 17 U.S.C. § 512 and your hosting provider takes the newsletter offline.
Your options: file a counter-notice under § 512(g), which puts the burden back on the claimant to sue — and which you can only file in good faith; negotiate a retroactive license; or accept the takedown. If you want to avoid this from the start, the sequence is: (1) check the copyright notice, (2) check for any license on the specific work, (3) check the permissions page, (4) if nothing, assume no permission and either link instead of copy or write a request. Sending a written permission request to the site's published contact address creates a paper trail and often gets a yes for small uses.
If you do get a takedown notice, don't ignore it. § 512 requires a response within the timeframe in the notice (commonly 10 business days for counter-notice), and failing to act can result in your account being terminated under repeat-infringer policies. This is administrative risk, not criminal risk. Misunderstanding that difference is why small operators get blindsided.
Terms of Service Can Prohibit What Copyright Might Allow
A terms of service agreement can restrict conduct even where the copyright owner's rights wouldn't. The two are independent: one is a property right (copyright), the other is a contract (ToS). If a site says "you may not use automated means to access the site," and you scrape it, the site can sue for breach of contract even if fair use would have covered the copying.
The enforceability question turns on notice and assent. Courts have enforced clickwrap terms (you clicked "I agree") more consistently than browsewrap (a link in the footer you never saw). The Ninth Circuit's Nguyen v. Barnes & Noble (2014) declined to enforce browsewrap terms where there was no evidence the user saw them; the Second Circuit has enforced terms where continued use after notice constituted assent. The practical rule: if you never clicked anything, and the terms weren't conspicuous, the site's leverage is weak but not zero.
For automated access, robots.txt is not the copyright permission file, but it is the technical signal most sites use to indicate what crawling they permit. Ignoring robots.txt doesn't create copyright liability by itself, but it's strong evidence in a ToS or trespass-to-chattels claim. The safest reading is robots.txt tells you the site's access rules; the copyright notice tells you the copying rules; both matter.
Quote Standards for Editorial and Research Use Vary by Publisher, Not by Law
There's no statutory word count for quotation, and any guide that gives you a fixed number ("500 words is fine") is describing an editorial custom, not a legal rule. The custom varies by institution. University presses and academic journals typically accept short block quotes with attribution; commercial publishers ask for written permission above a threshold they set internally. If you're publishing and want a defensible position, ask the publisher you're writing for what their quoting standard is, not what the law allows.
The reason for this: even where § 107 would shield a two-paragraph quote, a publisher with a house rule of "one paragraph" will ask you to cut it. The house rule reflects a risk-aversion policy, not a legal boundary. It also reflects a commercial reality — obtaining permission is cheaper than litigating fair use, and insurers price that in.
For research use, the same logic applies to document delivery and course reserves. US libraries operate under the "fair use" analysis and the CONTU Guidelines adopted in 1976, which set a working figure of five articles per year from the most recent five years of a single journal per requesting library. CONTU has no statutory force — it was a legislative compromise — but it's the de facto benchmark many libraries follow. That's a useful number to know when you're told "you can't download that many."
Check the Notice's Year, the Domain, and the Archive Before You Rely on Any of This
Two details at the bottom of a page change what you're looking at. First, a stale copyright year ("© 2019") on a live site tells you the site hasn't updated its footer, not that the content is unprotected — copyright lasts life plus 70 years for post-1978 works, and 95 years from publication for pre-1978 US works. Second, the domain matters: a 2018 site relabeled as "2026" with new branding may be the same copyright owner, or may be a domain transfer to a new entity. The Internet Archive's Wayback Machine lets you check what the page's copyright notice said in prior years, which is useful when you need to prove what license a specific article carried at the time you copied it.
If you're proving provenance for a piece of content, capture three things: the page URL, the access date, and a screenshot showing the copyright line and any license. That's the minimum paper trail if a dispute arises later. For commercial reuse, get the permission in a signed email or PDF, not in a text message.
One boundary before we finish. This is general educational information about copyright mechanics, not legal advice, and the outcome of any specific dispute depends on jurisdiction, registration status, and facts I can't see from here. If you have a live takedown, a licensing question with money on the line, or an infringement dispute, consult a qualified IP attorney licensed in the relevant jurisdiction — I'd also check with your country's copyright office (the US Copyright Office, the UK Intellectual Property Office, or the Canadian Intellectual Property Office) for current guidance. If you're facing a dispute involving threats or harassment, contact local emergency or crisis services rather than trying to resolve it through the site's copyright process.
FAQ
Is the copyright notice the same as permission to reuse?
No. The notice identifies the owner and year; it grants nothing. Permission comes from a license (Creative Commons, CC0, or a written grant) or from a statutory exception like fair use. If you can't find a license, the default under the Berne Convention is that all rights are reserved.
Can I screenshot a site for a review without permission?
Usually, if the use is genuinely editorial and limited to the review context — a full-screen shot illustrating your critique of that site carries a stronger fair use case than the same screenshot used decoratively. That's a defensible analysis, not a guarantee, because fair use is decided case-by-case on the four § 107 factors.
Does a site need a © symbol to be copyrighted?
No. In the US, notice stopped being required for works published on or after March 1, 1989. A page with no © is still protected. The notice does add evidentiary weight — 17 U.S.C. § 401(d) blocks an "innocent infringement" defense where a proper notice appears on the work.
Why did my republication of a short excerpt still get a takedown?
Because the fourth fair use factor — market substitution — is about whether your excerpt gives away what the original was selling, not about word count. A ten-word excerpt that answers the question the original article existed to answer can substitute. Length isn't the metric. Reciprocally, a longer excerpt in a scholarly critique that the original author expects may be defensible.
What's the maximum statutory damages I could face for one image?
$150,000 per work if the infringement is willful and the work was registered before infringement began, under 17 U.S.C. § 504(c)(2). The floor for ordinary statutory damages is $750 per work under § 504(c)(1). Registration is the trigger — § 412 bars statutory damages and fees for unregistered US works, which is why photographers register.
Does robots.txt give me permission to scrape a site?
No. robots.txt signals what crawling a site's operators permit technically. Ignoring it doesn't create copyright liability by itself, but it's evidence in a terms-of-service or trespass claim. Copying the scraped content is a separate copyright question — robots.txt doesn't displace the copyright notice or the terms of service.
Can I use a CC-licensed image without crediting the photographer?
Only if the license allows it. CC0 1.0 allows use without attribution — it's a waiver, not a license with conditions. CC BY 4.0 requires identifying the creator, naming the license, linking to it, and indicating changes, per the license's Section 3(a). "Photo from [site]" does not meet the CC BY attribution standard.
Which law applies if I quote a US site from outside the US?
Usually your local copyright law governs conduct you carry out in your own country, but the site's owner can enforce wherever the site is registered and where the harm occurs. The UK's fair dealing, Canada's fair dealing, and the EU's text-and-data-mining exception all differ from US fair use. Conflicting outcomes are common in cross-border disputes, so check the jurisdiction you'll be publishing in, not the one you're copying from.