Cloudflare has added a configuration option that lets a Worker serve its static assets from a URL path prefix rather than only from the site root, according to a changelog entry on the company's developer documentation site. The change means a project can expose its built files at an address like example.com/blog/ while keeping those files at the top level of its asset directory.

The changelog describes the previous behaviour plainly: to serve assets from /blog/, developers had to nest their files inside a matching blog directory in the build output. The new setting removes that requirement, so the directory layout of the build no longer has to mirror the public URL structure.

Two configuration surfaces are documented. In a Wrangler configuration file the option is written as assets.base_path, while in cloudflare.config.ts it appears as worker.assets.basePath. The changelog supplies example configurations for both, including a TypeScript example that defines a worker named "blog" with a compatibility date and an assets block containing basePath set to "/blog".

Version requirements are stated explicitly. The changelog says Wrangler 4.148.0 or later is needed, and that projects using cloudflare.config.ts require cf 1.0.0-beta.13 or later. The beta version number is worth noting for teams that pin tooling versions in CI, since the configuration-file route depends on a pre-1.0 release.

The changelog illustrates the routing behaviour with a concrete mapping. With the example configuration, a request to /blog/ serves dist/index.html, and a request to /blog/posts/post1 serves dist/posts/post1.html. Requests that fall outside /blog/ do not match the static assets and instead run the Worker, if the project has one.

That last detail matters for anyone running a Worker alongside static files. The prefix effectively scopes which requests are handled as asset lookups and which are passed through to Worker code, so the boundary between static and dynamic handling is now expressed in configuration rather than implied by directory nesting.

The changelog also flags several places where the full public path must be used. Paths written in _headers and _redirects files, and the run_worker_first setting, are all described as using the full public path including the /blog prefix. Developers migrating existing projects will need to check those files rather than assuming relative paths still apply.

There is an additional requirement for teams building with the Cloudflare Vite plugin. The changelog states that in that setup, Vite's own base option should also be set to /blog/ so the two configurations agree. The entry does not elaborate on what happens if the two are mismatched, so the consequence of an inconsistent setting is not documented in the supplied text.

For freelancers and developers maintaining client sites, the practical appeal is structural. Projects that previously carried a nested directory purely to satisfy URL routing can now keep a flat build output and express the public path in configuration. That is an editorial reading of the documented mechanism, not a claim the changelog makes about build tooling outcomes.

The change is also narrow in scope. It concerns where a Worker's static assets are served from, not how they are built, cached or secured. The changelog does not describe any change to asset handling beyond the path prefix, and it does not present the option as affecting performance or delivery behaviour.

Editorially, the most useful way to read this is as a migration task rather than a rewrite. The changelog presents the option as an alternative to nesting files in a matching directory; it does not state what happens to projects that already use the nested layout, so teams should confirm their own setup against the documented examples before changing anything.

The unresolved questions are mostly operational. The changelog does not say whether the option interacts with existing redirects in ways beyond the stated full-path rule, nor does it describe behaviour for projects that combine a prefix with other asset settings. Teams should verify their own configurations against the documented examples rather than assume edge cases are covered.

A sensible conclusion for this audience is to treat the release as a small but concrete configuration improvement, adoptable when a project's build output and public URL structure have drifted apart. The documented version floors and the Vite base requirement are the details most likely to trip up an upgrade, and they are stated clearly enough to check before deploying.