What are Open Graph tags and what do they do?
Open Graph is a set of meta tags in a page's head. The spec site ogp.me (checked 2026-11-07) says it lets any web page become a rich object in a social graph. Four properties are required: og:title (name), og:type (type), og:image (representative image) and og:url (canonical URL). Its most direct use is the title, description and image shown when someone shares your link in a chat app or on a social platform.
In other words, OG tags are designed for how a link looks when it is shared, not for where a search engine ranks it.
Do OG tags affect whether AI answers quote you?
There is no direct evidence. We read these documents (2026-11-07):
| Source | Does it say OG affects AI citation or ranking? |
|---|---|
| Google Search Central structured data gallery | Open Graph is not mentioned |
| OpenAI crawler documentation (Overview of OpenAI Crawlers) | The body text does not mention og: tags |
| Perplexity crawler documentation (Perplexity Crawlers) | The body text does not mention og: tags |
| Anthropic's crawler documentation | Not verified (we did not read it line by line this time) |
So the claim that adding og tags makes ChatGPT or Perplexity more likely to cite you has no source we can find, and we do not state it as fact.
One link does have a source: Google's documentation (Influencing your title links, checked 2026-11-07) lists the sources it uses to pick a title link automatically, and the og:title meta tag is among them, alongside the title element and h1. That is about title links in Google search results, not AI answers.
Should you still write them? Three reasons
1. The share card is the reader's first look
When someone drops your link into a chat or social post, the title and image on the card are your storefront. This rests on the ogp.me definition.
2. og:title is one of Google's fallback sources for title links
A clear title element matters more; keeping og:title consistent with it avoids the sources contradicting each other.
3. Cheap to do, but do not expect citations from it
We found no official statement on whether link cards in AI answer interfaces read og:image, so that is not verified. Treat OG as sharing experience, not as a GEO tactic.
What does a minimal version look like?
Below is a minimal version, with title and description matching the page itself and a 1200×630 image:
<meta property="og:type" content="website"> <meta property="og:url" content="https://example.com/"> <meta property="og:title" content="Product name: what it does in one line"> <meta property="og:description" content="Who it is for, what it does, how it is priced."> <meta property="og:image" content="https://example.com/og/home.png"> <meta property="og:image:width" content="1200"> <meta property="og:image:height" content="630"> <meta name="twitter:card" content="summary_large_image">
A real example: the home pages of our four sites
On 2026-11-07 we read the source of four home pages with curl, counting only the og tags written in the head:
| Site | og:title | og:description | og:image |
|---|---|---|---|
| mochistamp.com | Yes (Japanese) | Yes (Japanese) | Yes, 1200×630, /assets/og.png (opens with 200) |
| gammaslide.com | Yes (English) | Yes (English) | Yes, 1200×630, /og/home-en.jpg (opens with 200 via GET; a HEAD request returns 405) |
| kagentx.com | Yes (English) | Yes (English) | Missing: the home page source has no og:image |
| pagent.kagentx.com | Yes (English) | Yes (English) | Yes, 1200×630, /og/home.png (opens with 200) |
In short: all four sites write a title and description, and kagentx's own home page lacks og:image, which we will add. This data only shows whether the tags are present, not whether any AI used them.
How do you check your own pages?
curl -sL https://your-site.com/ | grep -i "og:"
The output should include at least og:title, og:description, og:image and og:url. The image URL must open directly, and 1200×630 is the suggested size. Write the tags into server-rendered HTML so they can be read even when a crawler does not run scripts; see Do AI crawlers render JavaScript?. For a full checklist see AI visibility audit checklist.
What really affects whether you get quoted is the page's text and structure; see Does structured data help AI search? and How to write content AI will quote: before and after.
FAQ
Do Open Graph tags affect AI search ranking?
We found no Google, OpenAI or Perplexity document saying they do. We did not read Anthropic's documentation line by line this time. Do not treat OG as a citation factor.
Does og:image show in ChatGPT link cards?
We found no official statement, so this is not verified.
Must og:title match the title element?
They need not be word for word, but the meaning should agree. Google uses og:title as one source when it generates title links automatically, and contradicting sources make the result inconsistent.
What size should og:image be?
Three of our four sites use 1200×630, and we suggest the same. That is our practice, not a requirement of any official spec.
What happens without og:image?
Most platforms show a text-only card when sharing. kagentx's home page lacks it right now, and we have no data on its effect on traffic.
Sources
- ogp.me · The Open Graph protocol (checked 2026-11-07)
- Google Search Central · Influencing your title links in Google Search (checked 2026-11-07)
- Google Search Central · Search Gallery (checked 2026-11-07)
- OpenAI · Overview of OpenAI Crawlers (checked 2026-11-07)
- Perplexity · Perplexity Crawlers (checked 2026-11-07)
- MochiStamp · Home page source, og tags (Read 2026-11-07)
- Gamma AI Slide · Home page source, og tags (Read 2026-11-07)
- kagentx · Home page source, og tags (Read 2026-11-07)
- Pagent · Home page source, og tags (Read 2026-11-07)
- kagentx evaluations: the data on the case study pages; each example gives its date.