HTML file · HTML to link · One page

One HTML file, one link. Drop it, or paste the code.

Whatever the file is called and whatever wrote it — you, a template, a chat — drop it and it is live over HTTPS at an address you can send. Or paste the markup straight in. No account, no repository, no build.

Drop a folder to start a new site

One HTML page of any name, a folder with an index.html, or documents — markdown, PDF, Word, Excel.

  • It opens as a page, not as a download

    A file attached to a message arrives as something to save and then open. A link opens in the tab the person already has, on a phone as easily as on a laptop.

  • Send the link once, change the page as often as you like

    Publish a corrected file to the same site and the address you already sent shows the new version, with every earlier one kept behind it.

  • Nothing counts the people who open it

    No visits, no bandwidth, no page requests — on any tier. A page that gets forwarded around costs exactly what it cost before it was.

When this is the wrong tool

  • ✕The page needs files you did not send. A page that loads an image, a stylesheet or a script from beside it on your disk needs those files published with it. Put them in a folder with the HTML, keeping the paths the page uses, and drop the folder instead of the file.
  • ✕It only worked because it ran inside a chat. A page that calls the model, or relies on libraries the chat's preview supplied, will be blank or broken anywhere else — this host included. The fix is in the page, not in where it is hosted.
  • ✕You want it to show up when people search. The default subdomain is served with a header asking search engines not to index it. A page meant to be found needs a custom domain, and that is a paid tier.
  • ✕It needs a server to work. A form is collected for you without a backend, and on the plans that include them a serverless function can answer an API route beside the page. PHP, a database or a page rendered per request still needs a host that runs a server.

Three steps, and none of them is a signup

Publishing needs no account at all. Signing in is what stops the site expiring and ties later publishes to the same address.

  • 1

    Take the file as it is

    Any name will do. A lone page.html is published as the site's front page, so the link opens on it without a file name at the end of the address.

  • 2

    Drop it, or paste it

    Drag the file onto the box at the top of this page, or paste the markup into it. The link comes back in seconds, before you have signed up for anything.

  • 3

    Keep the link, replace the page

    Sign in and the site stops expiring. Publish the next version to the same site and the address you sent stays the one that works.

Quota

Nothing counts your readers.

Files are served from the edge out of storage where egress costs nothing, so there is no traffic figure for your pages to pass on — and no tier here on which one appears. What is bounded is publishing, and the calls to your functions and to your own API, and those figures come from the same table the server enforces. Every tier, side by side →

2,000 files a publish, free

With an account and no card. A folder of pages and images meets the file count long before it meets the bytes.

10 MB a file, free

Per file rather than per site, so one large image is the thing to look at rather than the total.

0 versions kept

Each publish is its own version rather than an overwrite, and rolling back is a click. Paid tiers keep more.

  • ✕Never visits, bandwidth, page requests or build minutes. There is no build, and traffic to your pages is neither billed nor throttled.
  • ✓New or changed files only. Unchanged files are reused rather than uploaded again, so publishing the same site twice costs almost nothing.

Why an attachment or a cloud drive is not the same thing

An HTML file is only a page when something serves it over HTTP. Attached to an email it is a download; double-clicked, the browser opens it from disk, where module scripts and requests for the files beside it are refused — which is why a page that worked in a chat's preview can open blank. Cloud drives keep the file rather than serving it as a site: Google Drive did serve pages once, and switched that off in 2016.

Publishing it here is serving it. The file is put behind an HTTPS address on its own subdomain, sent with the content type a browser expects, and kept as a version you can return to. The people you send it to open a page, and you keep the one thing an attachment never gives back — the ability to change what they are looking at.

Questions

Is it really free, and for how long?

Publishing needs no account and costs nothing. A page published without one expires after a few days — the exact window is on the pricing page, read from the same table the server enforces. Signing in claims the site you already made, and it stops expiring.

Can I host an HTML file in Google Drive instead?

Not as a web page any more. Drive served pages from a public folder until Google switched the feature off at the end of August 2016; a file shared from Drive is now offered to preview or download, not as a site somebody browses.

Does the file have to be called index.html?

Not when it is the only file. A single HTML page is published as the site's front page whatever its name. In a folder of several files, the one called index.html is the front page and the others are served at their own paths.

Can I keep it private?

On a tier that allows it, yes: a password checked at the edge before any file is served, with the page kept out of shared caches while it is on — or a site only the people you invite can open, who sign in to see it.

Try it with the folder you already have.

No account and no card to publish. Sign in and the site stops expiring.

Share AI-generated pages and docs as a link · AI-generated HTML blank page when opened? The fix · Password protect a static site