htmlshare turns an html file into a public url. you drag a file onto the page, or send one post request, and you get back a link that opens in any browser. there is no account to create, no build step, no deploy, and no dashboard to configure first. the whole thing is designed to take less time than explaining what it does.

it exists because a specific kind of file kept having nowhere to go. ai tools — claude, chatgpt, coding agents — produce self-contained html constantly: reports, dashboards, one-pagers, charts, summaries. they are real, finished documents, but sharing one meant zipping it, emailing it, and asking someone to unzip and open it locally. static hosting solves that, at the cost of being far more work than the artifact is worth. htmlshare is the missing step in between.

uploads are stored on cloudflare r2 and served from cloudflare's edge network, so a page loads quickly wherever it is opened. we serve exactly the html that was uploaded: we never inject scripts, analytics, banners, or branding into a published page, and we never modify the markup. what you publish is what a visitor receives, byte for byte.

anonymous pages are free and are deleted seven days after publishing. that expiry is deliberate — it keeps abandoned pages from accumulating forever, and it keeps the free tier honest. signing in makes your pages permanent, lets you choose the url, and gives you a dashboard to manage what you have published. api keys do the same for programmatic access.

every upload runs through automated phishing and malware screening before it goes live, and every page carries a way to report abuse. we review reports and take down pages that violate the terms. beyond those safety checks we do not read, analyze, or sell the contents of what you upload — see the privacy policy for specifics.

the drag-and-drop page is one interface; the api is the other, and neither is an afterthought. there is a documented rest api, a published openapi specification, and an mcp server that lets claude, chatgpt, and other agents publish pages natively as a tool call. the llms.txt file describes when an agent should reach for htmlshare and how to call it. full details are in the api docs.

htmlshare is built and operated independently, not by a venture-backed company with a support org behind it. that means it is small and focused, and it also means questions go to a real person rather than a ticket queue. if something is broken, unclear, or missing, get in touch — support, abuse reports, and press enquiries all reach the same inbox.