Start with the right website

Open the exact article, guidance page, dataset description, or organization page that supports your statement and copy its canonical address.

Look for a named writer, an editorial group, or the organization responsible for the page. Do not automatically treat the site name as a personal author.

Record a publication or last-updated date only when the page clearly applies that date to the content. A footer copyright year is not the publication date of every page.

How AMA treats this source

Web pages are harder to normalize than journal articles because sites expose metadata inconsistently. A browser tab may show a shortened title, while the visible heading has the full title. An Open Graph author can refer to a social account rather than the person who wrote the page. Automated extraction is a starting point, not proof.

AMA website references distinguish the page title from the name of the overall site. They may also identify the authoring organization and the dates relevant to publication, update, and access. If the page changes over time, the access date helps describe the version you consulted.

A URL should take the reader as close as practical to the cited content. Remove tracking parameters when they are not needed to reach the page, but do not shorten the address so aggressively that it stops working. When a web page is actually the landing page for a journal article or report, cite the underlying article or report type instead.

Fields to verify before copying

Automatic metadata is useful, but it is not a substitute for the source. Compare these fields with the item you actually read.

Field	What to check
Author or organization	Use the responsible person or body shown on the page; do not infer one from the domain.
Page title	Use the title of the individual page, not just the browser tab or site-wide slogan.
Site name	Record the larger website or publisher when it differs from the page title.
Dates	Confirm publication or update date and use an access date when the applicable guidance requires it.
URL	Use the direct, working page address without disposable campaign parameters.

Worked AMA example

Source being cited: The AMA Manual of Style website, published by Oxford University Press, accessed September 2, 2026.
Reference-list entry: 1. AMA Manual of Style. Oxford University Press. Accessed September 2, 2026. https://www.amamanualofstyle.com/
In-text use: The first mention points to superscript reference 1: ¹

This example has no named page author or publication date, so it does not invent either field. A dated article page would need its own details.

A reliable workflow

1Identify the underlying website and collect the most stable identifier available.
2Generate the reference, then compare every imported field with the publisher record or the source itself.
3Place the source in the reference list according to the order in which it first appears in your writing.
4Use that assigned number again whenever you cite the same work; do not create a new reference-list entry.
5Check the final manuscript against the current instructions from your course, journal, or publisher.

Common mistakes to catch

Citing Google or a database result

Cite the page you opened and used, not the search service that helped you find it.

Using the footer year as the page date

A copyright notice is site-wide and usually does not establish when the cited content was published or revised.

Dropping the page title

The page and website are separate elements. A reader needs the individual page title to understand what you used.

Leaving tracking parameters in place

Campaign parameters make references noisy and can expose how you reached the page. Keep only what is necessary for a stable destination.