Resolve a relative URL against a base URL

Provide an absolute base URL and the reference found in a document, feed, stylesheet, manifest, or API response. The resolver applies standard URL behavior to path segments, trailing slashes, queries, fragments, ports, and origin changes, returning one normalized absolute URL without requesting the destination or inspecting any remote content. It is fast and deterministic.

Start with the complete URL that supplies context

A relative reference has meaning only in relation to another URL. Enter that context in base_url as a complete absolute URL, including its scheme and host when the scheme uses a host. The path matters as much as the domain. For example, a base ending in <code>/docs/page.html</code> treats a plain reference such as <code>image.png</code> as a sibling of the page, while a base ending in <code>/docs/</code> treats it as a child of that directory. The resolver does not guess whether a path represents a file or folder; it follows the syntax, especially the final slash. It accepts standard absolute URL forms understood by the browser URL implementation. A value such as <code>/docs/page.html</code> is not sufficient as a base because it lacks an absolute scheme and origin context. Invalid or relative bases produce an input error instead of a plausible-looking answer. This strictness makes mistakes visible before resolved links enter a crawler, import job, redirect rule, or stored dataset.

Understand how each kind of reference changes the base

The relative_reference field can contain more than a simple filename. A leading slash replaces the base path while retaining the base scheme and authority. Two leading slashes create a network-path reference, allowing the host to change while inheriting the scheme. Dot segments such as <code>../</code> move up path levels and <code>./</code> stays at the current level; normalization removes those navigation segments from the final URL. A query-only reference beginning with <code>?</code> keeps the path and replaces the query. A fragment-only reference beginning with <code>#</code> keeps the existing path and query while changing the fragment. A fully absolute reference replaces the base altogether, as standard resolution requires. Empty references are rejected by this capability so accidental missing form values are not mistaken for deliberate reuse of the base. Percent-encoded characters remain encoded according to URL serialization rules, while host names, default ports, and other components may be normalized into the canonical form produced by the standard URL parser.

Use the serialized result without triggering a network request

The result contains absolute_url, the normalized serialized URL after resolution. No DNS lookup, HTTP request, redirect follow, page download, or availability check occurs. That separation is useful when you need deterministic transformation rather than a live-site test: the same two inputs always return the same output, even when the host does not exist or the resource is private. Use the result to normalize links extracted from HTML, prepare crawl queues, expand references in feeds, validate fixtures, build sitemap entries, or compare URLs consistently before deduplication. Remember that syntactic resolution does not prove that the destination is safe, reachable, public, or authorized. If untrusted users supply the values, apply your application’s scheme, host, and network-access policy before fetching the returned URL. The API and browser runner use the same pure algorithm, so interactive checks and automated calls agree. Each API request costs $0.002; browser execution on this page is free and keeps the calculation local.