CardCite is a Chrome extension that builds debate citations. Open it on any webpage or PDF and it reads the page, fills in the publisher, date, author, title, source description, and URL, and assembles the citation in your team's exact format.
No hunting through the page for a date, no looking up who the publisher is, no retyping brackets. One click, then paste into your doc with fonts, sizes, underlines, and italics intact.
The popup on a Reuters article: every field filled from the page, the finished citation previewed at the top, each value editable before you copy.
The popup on a Congressional Research Service report: publisher, date, both authors, title, URL, and a source description, all read from the page.
The popup is a small editable form. Everything you need for the citation arrives filled in and ready to check, pulled from the page rather than typed by hand.
One uncut take: open the popup on an article, look over the fields, copy, and paste into a doc with the formatting intact.
Highlight a passage before opening CardCite and it carries over, shown in a small preview under the citation. When you copy, the quote pastes along with the citation: the citation in your format, the quoted text in plain 12pt Times New Roman.
A citation is only useful if it holds up when someone checks it. Most of CardCite's engineering effort goes into one problem: never showing you a confidently wrong value.
CardCite doesn't lean on any single part of a page. For each field it gathers the signals a page carries, the metadata in its source code alongside the text a reader actually sees, and confirms them against each other before a value is used. The specifics vary by field: each author candidate is screened against more than 83,000 words that appear in ordinary English but never in real names, the publisher is looked up in a built-in list of more than 20,000 domains, the publication date is told apart from the day the page was last edited, and the title is stripped of the site name and boilerplate appended to it. All of it has been refined against hundreds of real articles and documents, so the value that reaches your citation is the one the page genuinely supports.
Measured against a hand-checked sample of real articles and PDFs.
A short account of how the benchmark is built and run goes here: how the sample was chosen, what counts as correct, and how misses break down between blank and wrong. Full field-by-field results are coming after the complete test.
Dates are shown in many formats: machine timestamps, written dates, or even a relative label like "3 days ago". CardCite is built to correctly interpret all of them, drawing from both the page's structured data and its visible text. When a page shows a date it was updated on, that updated date is cited by default and "Updated" is added to the citation. If only the publication date is shown, or if no date is shown on the page, the published date is used. If you prefer the publication date to always be cited, you can set that in settings. When no date appears in the page's source code, CardCite uses the one shown on the page with a note to verify it.
Byline areas are full of things that aren't names: staff labels, photo credits, sponsor tags. Every author candidate is screened against a list of more than 83,000 words that appear in ordinary English but not in real names, compiled from large word-frequency corpora and cross-checked against name datasets so that real but unusual surnames still pass.
Junk that fails the screen is dropped from the byline, and any real name left is kept and flagged for you to confirm. A byline can also carry real people who aren't the writer: when someone is credited under a production role, with a label like "Photos by", "Illustration by", or "Translated by", CardCite reads it as a credit rather than a byline and keeps that name out of the author field.
What's on the page, and what ends up in the citation.
Every citation names the outlet it came from. CardCite reads the publisher from the page and checks it against a built-in list of over 20,000 domains, so an article at nytimes.com is cited as The New York Times, in the wording and capitalization the outlet itself uses.
The list also fixes what sites write in their own code, where publisher names often arrive lowercase or shortened. For anything it doesn't cover, CardCite builds a clean name from the address, so the field is never left for you to fill in from scratch.
The address you're reading at, and the publisher name the citation gets.
Two kinds of messages appear under fields, and they mean different things. Amber notes mark values that came from a weaker signal and deserve a look before you rely on them. Quiet notes don't question the value; they tell you something about where it came from that a careful researcher would want to know.
Small nudges cover the rest, like a warning when a scraped publisher name arrives with no capital letters.
An amber note: worth double checking the field.
A quiet note gives added context about where a result came from or why it was selected. It doesn't mean anything is wrong, it just explains the choice, such as why an updated date was used, or where the exact day came from when the page prints only a month and year.
PDFs are opened with the same engine Chrome itself uses to display them, so modern, compressed, and encrypted files all read correctly, with a lighter built-in parser as the fallback for damaged files. Two sources are read together: the metadata the file records about itself, and the positioned text on the page, every fragment with its size and location. This helps tell a title apart from a byline, and a byline from body text.
The same corroboration applies. A date is only trusted when the cover pages support it, because report bodies are full of date-shaped noise, and a re-hosted copy that stamps its own regeneration date gets a blank date instead of a wrong one. Recorded metadata is checked against what's printed on the cover before either is believed.
Every change to the detection code runs against a regression suite of hundreds of real saved pages and PDFs before it ships.
When CardCite recognizes the site you're on, it fills in a short description of the outlet, drawn from a built-in database of 600+ sources covering news outlets, think tanks, journals, and government bodies, with new sources added over time for upcoming resolutions.
For everything else, write the description once and save it. It's tied to that site and fills in automatically the next time you cite anything there. Every description is fully editable, including the built-in ones: change the wording on any of the 600+ sources and CardCite saves your version and uses it from then on, in place of the original.
Saved descriptions sync with Chrome across your devices and can be edited or removed anytime in the Library.
Write a description once, or edit the built-in one, then save it for the site. Next time you cite anything there, it's already filled in.
Every citation you save from the popup lands in the Library, kept exactly as it was formatted when you saved it. It's built for two things in particular.
Going back to a source before you use it. Each entry keeps its link and its details, so when you want to reopen an article and read it properly before pulling a quote from it, the page is a click away instead of something you have to track down again.
Getting citations into a document. The CardCite add-on for Google Docs™ opens your saved library in a sidebar beside what you're writing, so you can insert citations as you go. When the same source comes up again later, it's still in the sidebar rather than something to scroll back and copy.
Save any citation straight from the popup. Each entry keeps the exact formatting it had, fonts, styling, and field order included.
Find past citations by publisher, title, author, date, or URL, and by any note you've attached to them.
Open a saved citation and change any field. It rebuilds in your current format settings, and a delete can be undone.
A separate tab lists every description you've written, searchable and editable, with the built-in one restored if you delete yours.
CardCite is built to be deeply customizable. Start from a preset: the CardCite default, MLA, APA, or Chicago, then change anything you like, and your changes are kept as a Custom preset you can return to. Settings sync through Chrome to every computer you sign into.
The settings page. Drag a field to reorder it; open its Format drawer for enclosure, punctuation, size, styling, and prefixes. A live preview updates beside it as you work.
Each field carries its own brackets or parentheses, ending punctuation, font size, bold, italic, and underline, and whether it starts a new line. Between them, almost any citation format can be matched exactly. CardCite also allows prefixes to be added automatically, like "Written by" ahead of the author names or "According to" ahead of the publisher. Choose from a few presets on each field, or write your own.
Choose whether a page that shows both a published and an updated date cites the update or the original.
A compact mode opens straight to the finished citation, ready to copy. You can also move the citation block below the fields, close the popup automatically after copying, and change the keyboard shortcut (Alt + C by default).
One toggle themes the popup, the settings page, and the Library.
Compact mode and dark mode together: the popup opens straight to the finished citation, ready to copy.
