HTML file mode
Route: Spread → HTML file tab
Drop in a .html file and Spread rebuilds it exactly as it would a fetched page.
No network request is made at all.
This is not the fallback — for apps, it is the main path
Two very common cases cannot be reached by pasting a URL, and both are solved here:
JavaScript applications. React, Vue, Angular and friends render in the browser. Fetched over HTTP they return an empty shell — often literally "You need to enable JavaScript to run this app". No amount of AI can rebuild a page from that, so Spread detects it and stops.
Anything behind a login. A dashboard, an admin panel, your own product. The interesting screens are almost always the signed-in ones.
How to save a rendered page
- Open the page in your browser and sign in if needed.
- Wait for it to finish loading.
- Save As → Web Page, HTML only (Chrome, Edge, Firefox all have this), or
open DevTools and copy
document.documentElement.outerHTML. - Upload the file to Spread.
Why we do it this way
Tostada could offer to log in for you. It deliberately does not.
Your application credentials are not an API key: unscoped, hard to revoke, often an admin account. Taking custody of them would make us responsible for something we have no business holding, and it would still fail on SSO, MFA and captcha — which is most of the enterprise world.
Saving the page yourself means your credentials never leave your browser, and single sign-on works for the simple reason that you already got through it.
Differences from URL mode
- No "View original" link in the report — there is no URL to point at.
- Relative image paths in the saved file may not resolve. Spread reports any asset it could not render rather than quietly replacing it with a grey box.
- Everything else — the checkpoint, the legacy choices, the estimate, the report, the slider, the download — is identical.