demo · v134
Single-page app shell — measured savings
The chromestatus explainer’s named target: tiny SPA shells. The HTML is a few hundred bytes — just <div id="root"> and a script tag — but Chrome still runs speculative HTML pre-parsing looking for <link rel="stylesheet">, <script src>, <img>. With Document-Policy: expect-no-linked-resources, the pre-parser can be skipped and the document goes straight to JS. Pick a shell size below and watch the modelled critical path.
probing Document-Policy support…
Pick a shell, compare paths
without expect-no-linked-resources
HTML download0 ms
speculative parse0 ms
JS executes0 ms
(no header)
with expect-no-linked-resources
HTML download0 ms
speculative parse0 ms
JS executes0 ms
Document-Policy: expect-no-linked-resources
What it isn’t
This is a hint, not a sandbox. If the document later contains a <link> or <script src> the resource still loads. The page just promises “you won’t find anything by pre-parsing,” letting Chrome skip a doomed pass on shells where every byte counts.