HTMLRadar · About

About HTMLRadar.

Updated September 2026

HTMLRadar is an open-source tool for sharing an HTML deck, brief, or proposal as a tracked link, and seeing who opened it, which sections they read, and for how long. It is AGPL-3.0, built by the HTMLRadar team, and the whole thing runs on your own Cloudflare and Supabase accounts if you would rather host it yourself.

The facts, in one place

Who builds it
The HTMLRadar team. Reach us at [email protected], or open an issue on GitHub.
Since
The first commit landed on 13 May 2026.
Licence
AGPL-3.0-or-later, for the tracker, the proxy worker, the database schema, and the web app. A commercial licence exists for anyone who cannot accept the AGPL's terms.
Source code
github.com/htmlradar/htmlradar — the whole product, including the parts that run the hosted service.
Package
The MCP server is published on npm as htmlradar-mcp, and listed in the Model Context Protocol registry as com.htmlradar/share. A remote connector runs at mcp.htmlradar.com/mcp. See /mcp.
How the sender learns who read it
A tracked link asks the reader for an email address before it opens the document. That gate is on by default and can be switched off per link; with it off, a read is recorded as a visit rather than a named person.
Where documents are served
Tracked links are served from a second domain, htmlradar.page, so a recipient's browsing of a shared document never shares cookies with the marketing site or the dashboard. Recipient documents are never indexed. The reasoning is in the privacy policy.
How it runs
Cloudflare Pages, Workers and R2 for serving and storage; Supabase for the database and sign-in. The same stack self-hosts in your own accounts — see /self-hosted.
What it costs
Free for two tracked links, then $15 a month or $150 a year for unlimited links. The software is free to self-host; you pay your own infrastructure providers directly. Full pricing.

Why it exists

More of the documents that matter are HTML now, because that is what an LLM hands you, but the tools for learning whether one was read were all built around uploading a file. The argument in full is on the why page, and the engineering is written up on the blog.

Corrections

If anything on this page is wrong or has gone out of date, tell us at [email protected] and we will correct it and say what changed.