Article Schema Generator
Article markup tells a search engine who wrote a page, when it came out and when it last changed, and which images belong to it. This generator writes it as JSON-LD, with as many authors and images as the page has and every date in ISO 8601 with the right offset, then checks it against rules taken from Google's Article documentation. Everything you enter stays in this browser tab.
Google's Article page covers all three; pick the one that describes the page best.
The title of the article as the page shows it. Google sets no length limit, but long titles may be cut on some devices.
Google's page asks for a date and time and recommends time zone information. Date only writes the day alone, which schema.org also accepts.
Starts on this device's time zone, and each date then gets the offset that zone uses on its day, daylight saving time included. A fixed offset is written exactly as chosen.
Pick a fixed offset when the update was made in another place than this device.
Written as an Organization. publisher is a schema.org property; Google's Article page does not list it.
Article, NewsArticle or BlogPosting JSON-LD, or the HTML of a page that holds it. The button replaces everything in the form above.
Limits: Up to 10 images, 20 authors and 1 MB of text in each box. Larger input is refused with a message.
The checks are rules this page took from Google's Article documentation (last updated 8 September 2026), read on 22 October 2026. Run the published page through the Rich Results Test and the Schema Markup Validator as well, since they decide and this page does not.
Results
- Type
- -
- Images
- -
- Authors
- -
- Findings
- -
| Level | Where | Finding, why it matters and the fix |
|---|
JSON-LD
Script element for the page
How to use the Article Schema Generator
- Pick Article, NewsArticle or BlogPosting, then write the headline and the full address of each image.
- Enter both dates, and change their time zones if they happened somewhere other than this device's zone.
- Add each author with a type and a page about them, or import markup you already have to fill the form.
- Fix the findings, copy the script element into the page, and test the published page with the Rich Results Test.
Worked example
A blog post went up in New York late on 7 March 2026 and was updated two days later, after the clocks had moved forward. Press Load the example to import its markup into the form. The two dates keep their offsets, -05:00 and -04:00, the second image has a relative address, and the second author has no address yet.
{
"@context": "https://schema.org",
"@type": "BlogPosting",
"headline": "How to repot a fern without hurting the roots",
"image": ["https://www.example.com/photos/fern-16x9.jpg", "/photos/fern-4x3.jpg"],
"datePublished": "2026-03-07T22:15:00-05:00",
"dateModified": "2026-03-09T08:00:00-04:00",
"author": [
{ "@type": "Person", "name": "Willow Lane", "url": "https://www.example.com/staff/willow-lane" },
{ "@type": "Person", "name": "Sam Reed" }
],
"publisher": { "@type": "Organization", "name": "Fern Weekly", "url": "https://www.example.com/" }
}
- Type
- BlogPosting
- Images
- 2
- Authors
- 2
- Findings
- 6
How to read the results
- Type, Images and Authors
- The type the markup uses and how many image and author rows went into it, empty rows left out.
- Findings
- Recommended means a property Google's Article page recommends is missing, such as an author's address. A warning is a value of the wrong shape, like a relative image address or a time that does not exist. A note says how a date was read, or points out something to check by hand.
- Where
- The part of the form a finding is about: Headline, Image 2, Date published, Author 1, address, Publisher. Import marks what an import could not keep, and the type name means the markup as a whole.
- JSON-LD and Script element for the page
- The markup with two space indents, then the same text inside a script element of type application/ld+json, ready for the head or body of the page. Copy takes all of it.
How this tool works and its limitations
The form's values are trimmed, empty rows and empty fields are left out, and the rest is written in the order Google's page lists the properties, with each author typed as Person or Organization. A date and time is paired with an offset: a fixed one as chosen, or for this device's zone the offset the browser's time zone data gives for that day, found with Intl.DateTimeFormat. A skipped time moves forward by the length of the gap and a repeated time takes the earlier offset, the rule JavaScript's Temporal API calls compatible. The import reads JSON-LD, or a page's application/ld+json scripts through an inert DOMParser document, and takes the first Article, NewsArticle or BlogPosting it finds. One date with a time and the other without cannot both be kept as written, since the form writes both dates one way: the date alone gets 00:00 at the other date's offset, a time the import notes as invented.
The markup then goes through the Article rules the JSON-LD Generator uses: the seven recommended properties, ISO 8601 dates, full addresses, and author entries with a type and a single name. The form adds notes on one or two images, repeated images or authors, and the publisher. Output is escaped so that no text typed in a field can close the script element. These limits apply:
- The rules are this page's reading of Google's Article documentation, last updated 8 September 2026 and read on 22 October 2026; the Rich Results Test is what decides for Google.
- The form writes headline, image, the two dates, author and publisher. Image sizes as ImageObject, jobTitle, mainEntityOfPage and other properties have to be added by hand, and an import names each one it left out.
- The checks see addresses as text: they cannot open an image to measure it or tell whether Google can crawl it, and cannot compare the markup with the visible page.
- The fixed offsets cover those in use today, from UTC-12:00 to UTC+14:00. An imported offset outside that list is shown as the same moment at UTC.
- The form takes up to 10 images and 20 authors, and each box up to 1 MB of text; more is refused with a message.
Features
- Image and author rows with named Up, Down and Remove buttons that work from the keyboard.
- Dates written with the offset each one had in the chosen time zone, so daylight saving time is handled.
- An import that reads existing markup, or a whole page, and lists what it could not keep.
- Findings named by the part of the form they concern, with the reason and the fix.
What Google's Article documentation asks for
Google's Article documentation, last updated 8 September 2026, covers Article, NewsArticle and BlogPosting and says plainly that there are no required properties. It recommends author, author.name, author.url, datePublished, dateModified, headline and image. Its author advice is specific: list every author the page shows, each in an entry of their own, typed Person or Organization, with nothing but the name in the name. The page says the markup can help Google understand the article; it promises no particular result.
Dates, offsets and daylight saving time
Both dates are written in ISO 8601, and Google recommends including time zone information; without it, Google falls back on the time zone Googlebot uses, which its Googlebot page gives as Pacific Time when crawling from the US. The form starts on this device's zone and finds each date's offset for its own day. An article published in New York on 7 March 2026 is written with -05:00, and one updated on 9 March with -04:00, because the clocks moved forward on 8 March. A finding explains what happens to a time skipped or repeated by such a change. Date only writes YYYY-MM-DD, which schema.org's Date type allows.
Images and authors that match the page
Google asks for image addresses it can crawl and index, images that represent the article rather than a logo, and, for best results, several high resolution images of at least 50,000 pixels in the ratios 16x9, 4x3 and 1x1. Each author's address should lead to a page that identifies them; sameAs works too. None of this shows in the markup alone, so look at the published page as a reader would. Google's general guidelines also ask that markup describe content people can see, and they warn that markup which passes every check may still not appear in results.
Frequently asked questions
Which Article properties are required?
None, according to Google's Article page. It recommends seven: author, author.name, author.url, datePublished, dateModified, headline and image. This page reports a missing one as Recommended and never as an error.
Should I choose Article, NewsArticle or BlogPosting?
Choose the type that describes the page: schema.org defines NewsArticle as an article that reports news, and BlogPosting as a blog post. Google's page treats the three alike, and so do the rules here.
What happens if a date has no time zone?
Google says it uses the time zone of Googlebot instead, so 09:00 written without an offset can be read hours away from when you meant it. This page always writes an offset unless you choose Date only.
Why do my two dates have different offsets?
Because a daylight saving change fell between them in the chosen time zone. London is at +00:00 in winter and +01:00 in summer, so a February date and a June update differ by an hour, which is correct.
Can I give a date without a time?
Yes. Choose Date only and the markup holds YYYY-MM-DD, such as 2026-03-07, with no time and no offset. A note in the findings repeats that Google's page asks for a date with its time.
Do I need a publisher?
Google's Article page does not list publisher, and this page adds no rule for it. It is a schema.org property of every article, and Google's author advice names it as the place for the publisher's name. The form writes it as an Organization when you fill it in.
Last updated .