Enter up to three navigation names and URLs, then select Generate Schema. Complete pairs become a BreadcrumbList JSON-LD script block with consecutive positions; incomplete pairs are skipped.
The Breadcrumb Schema Generator creates a BreadcrumbList JSON-LD script block from up to three navigation steps. Enter a name and address for each step you want to include, then select Generate Schema. The output appears in the textarea below the controls. A pair is included only when both supplied strings are non-empty, and included entries receive consecutive positions beginning at one. Generation formats your supplied trail; it does not crawl a website, infer a category hierarchy or publish new navigation links.
A breadcrumb trail describes how a page fits within an understandable navigation path. Begin with the site's actual content structure rather than assuming every slash-delimited URL component belongs in the trail. For an example sequence such as Home, Guides and a specific guide, review both the labels and their intended public destinations. Copy the complete generated block only after that review. If you change a name or URL, select Generate Schema again because previously generated output is not automatically updated when fields are edited.
Each name field is intended to become the name of a ListItem. A simple sample might use Home, Guides and Rainwater Log to identify meaningful levels leading toward a page. The labels should correspond to the actual navigation context and make sense to a reader, not be a list of loosely related keywords. The form cannot check whether a category exists or whether its visible label agrees with the term you supplied.
Keep the trail specific to the page being described. A generic sequence copied to unrelated routes can misrepresent how the site is organized even when all values are strings and every URL looks plausible. If the information architecture changes, review the trail and the visible navigation together. Structured data does not create missing category pages, repair a confusing menu or establish a relationship that the site's content and navigation do not actually support.
The intended helper includes a row only when both the name and URL are non-empty. There are three rows, with the later rows labeled optional. A missing partner means that row would be skipped by the configured object-building logic. Whitespace-only values are not comprehensively trimmed and validated, so a visually blank field can still behave as non-empty. Review each pair deliberately instead of treating an accepted string as a verified navigation entry.
Use the actual intended address for each level. A label pointing at the wrong category is still wrong even if the final JSON syntax is valid. The mapping has no network lookup, address normalization or redirect check. Keep the reviewed trail with the page inventory so later moves can be updated consistently. A generator cannot decide whether your chosen link is public, appropriate or the preferred route merely from the fact that a name was entered beside it.
The configured position counter starts at one and advances only when a complete pair is added. If the first row is incomplete and the second is complete, that included second row receives position one. Positions therefore describe the sequence of included items, not the original form row numbers. This is a useful distinction when reviewing a proposed trail, because a skipped parent can leave an unexpectedly short list without an obvious position gap.
For a two-item sample, the intended object has ListItems at positions one and two in the order added. The tool does not sort them alphabetically or determine which address should precede another. Check the relationship yourself: a shuffled Home and category sequence can still have perfectly consecutive numeric positions. A sequence number records order, but it does not validate that the order represents a sensible path toward the content or matches the reader's navigation options.
The interface supplies at most three ordinary steps. A deeper trail needs a workflow that can represent the full intended list rather than omitting meaningful context solely because the form has no additional row. Conversely, do not invent extra steps just to fill all fields. The configured helper accepts only rows with both a name and URL; it does not offer an independently configurable final item with a different field requirement or a complete breadcrumb data editor.
Consult the applicable structured-data consumer guidance when the intended trail needs a representation beyond these controls. The form's visible layout is not the entire BreadcrumbList vocabulary, and it does not decide every implementation convention for your site. Keep a clear reviewed path and use suitable metadata tooling to express it accurately. A smaller correct trail is not automatically equivalent to any longer one, so explain any omitted or simplified relationship rather than silently treating all alternatives as interchangeable.
The intended JSON-LD object does not add a visible breadcrumb component to your page. It describes a list; it does not create links in the interface, add separators or style a navigation bar. If users need a visible trail, build and review that component through the site's normal design and template system. Avoid reporting a structured-data draft as if it had already improved visible navigation or made every category reachable from the current page.
Likewise, a visible breadcrumb can exist without this tool knowing about it. Compare the metadata values with the actual labels and destinations rather than assuming they remain synchronized after template changes. A shared component may rename a category while an old JSON block continues to use the previous name. Keep the trail's source of truth maintained so later edits do not create two conflicting descriptions of where the page belongs within the site.
The output contains an application/ld+json script element, formatted JSON and actual line breaks around the object. The generator uses JSON.stringify for string quoting and escapes less-than characters inside the serialized data. This lets a value containing a script-closing sequence remain part of its JSON string instead of becoming a separate HTML element when the block is embedded. Escaped characters still represent the original text after JSON parsing; they are not a reason to replace meaningful wording or edit the output by hand.
Copy the complete block from the output textarea rather than only a few visible property lines. Review the data values first, then inspect the JSON and its placement in the destination template as separate checks. A syntactically valid object does not establish factual accuracy, correct page association or a consumer's acceptance of every property. The form does not deploy the block or run an external validator. After integration, inspect the actual served route for the expected object and any duplicate or conflicting metadata from other site components.
Save the chosen path with the page revision and review it when categories are renamed, removed or moved. An old trail can remain syntactically plausible while leading to outdated addresses or describing a relationship that no longer exists. Keep any URL migration decisions with the metadata update, because generating a new object does not maintain redirects, rewrite internal links or create replacement category content for the old destinations.
The useful outcome is a truthful navigation description backed by the actual site structure. Generate a small known trail and compare the output's names, included entries, positions and URLs with the reviewed map before using it on more pages. Recheck the public route after publishing so a correct draft is not confused with a successful deployment. The generator does not promise a particular search display or rank unrelated pages. Keep the visible navigation and the structured description consistent, especially when a category moves, a label changes or an intermediate route is removed.
Enter the desired name-and-URL pairs and select Generate Schema. Review the output, including the included rows and positions, then copy the complete block.
Up to three name-and-URL pairs are available in the current interface.
The intended helper includes a row only when both values are non-empty. It does not comprehensively validate whitespace-only input.
No. The intended counter starts at one and advances for included items only.
No. You must supply a reviewed navigation relationship rather than expect it to crawl menus or interpret path segments.
No. A visible navigation component is a separate page-template feature.
No. The configured mapping copies supplied addresses without a live lookup.
No. Consecutive positions can still describe the wrong order or unrelated pages.
Review JSON syntax, the complete script block, included items, consecutive positions and the actual served page against the intended navigation trail.
No. Correct data and implementation do not guarantee a particular display or ranking outcome.