What Open Graph tags do
Open Graph metadata gives crawlers a structured title, description, image, canonical page URL, content type, and site name. A platform that receives a shared link can fetch the public HTML and use those values to build a rich attachment. Without suitable metadata, it may guess from the document title, visible text, or images, producing an inconsistent result.
The generator creates og: meta elements from your input. It does not publish the page or force any platform to use a specific design. Each service controls its layout, crop, moderation, truncation, and cache. The built-in preview is therefore a composition aid, not an exact reproduction of every sharing surface.
How to generate Open Graph markup
- Enter a concise title and a description that accurately summarize the destination.
- Supply the page's absolute canonical HTTPS URL and select the appropriate object type.
- Add a public absolute image URL, plus its dimensions and alternative text where available.
- Review the preview for cropping, repetition, and readability.
- Copy the tags into the server-rendered head, deploy, and verify the live response.
Use one coherent set of tags per page. A single global image and description may be acceptable as a fallback, but articles, products, tools, and events should describe their own content. In application frameworks, generate metadata from the same canonical record used to render the page so stale fields do not diverge.
Choose strong field values
Title, description, URL, and type
Write the title for a person deciding whether to open the link, with the distinguishing subject near the beginning. Use the description to add concrete context rather than repeat the title. The og:url value should be the clean canonical address without tracking parameters. Select a specific type such as article when it genuinely matches; use website for an ordinary web page.
Image and alternative text
Provide a stable HTTPS asset that responds without cookies, authentication, or expiring signatures. Use a composition that remains understandable under different crops, keeping faces, product details, and key typography away from edges. Declare accurate dimensions when known. Alternative text should convey the image's meaning rather than repeat promotional copy.
Useful publishing scenarios
A news site can populate article tags from its editorial system, including a unique lead image. A store can describe the canonical product rather than a filtered or campaign URL. A documentation team can give release notes a distinct title and graphic, while a small business can establish a reliable default image for pages that do not have their own artwork.
Open Graph can coexist with standard SEO metadata and Twitter Card fields. They serve overlapping consumers but should not contradict one another. Keep the HTML title, canonical link, structured data, Open Graph fields, and visible page content aligned. Metadata is not a place to advertise claims that a visitor will not find after opening the page.
Debugging missing or stale previews
Fetch the deployed page as a signed-out visitor and view the original HTML response. Confirm the tags are in the head before JavaScript executes, the image returns a successful image response, redirects terminate cleanly, and robots or firewall rules do not block the crawler. Duplicate tags from plugins or nested templates can make selection unpredictable.
Sharing services cache scraped values, sometimes for much longer than a browser cache. Correct the live source first, then use the platform's current inspection or rescrape workflow when one is available. Changing query parameters merely to evade a cache can fragment analytics and canonical signals, so preserve one authoritative URL.
Privacy and local generation
The markup and preview are generated in your browser, so draft text does not need to be uploaded to PagesTools. After publication, the metadata is deliberately public: crawlers and anyone viewing source can read it, and the image host receives crawler requests. Do not include private launch details, personal data, internal URLs, access tokens, or signed image addresses.