supermemoryai / preprint
PublicA new substrate for browser agents: files, actions, diffs, logs, and artifacts.
Preprint is a tool that bridges AI assistants and real web browsers by projecting browser content as simple text files. When you open a website, it creates readable markdown documents showing the page content. To interact with the page, an AI simply writes a plain English command into the file—like 'click the search button'—and preprint translates that into real browser actions. Everything stays synchronized automatically, so the text files always reflect what's actually on screen. It supports your existing Chrome profile with your logins and cookies, can capture screenshots and recordings, and works with popular AI coding assistants.
How It Works
You need your AI assistant to look up information, fill out forms, or click through websites—but you don't want to learn complicated automation tools.
One simple command downloads and sets up everything you need. It finds Chrome on your computer automatically.
You tell preprint to open any website—like news.ycombinator.com—and it launches a real Chrome window showing that page.
Preprint creates simple text files that show you exactly what's on the page. You can read them like any document to see the page content.
The AI can click any element it sees on the page by referring to it by name
The AI can type text into input fields, search boxes, or text areas
The AI can capture pictures of the page at any moment
Within about a second, Chrome does exactly what was requested. The text file updates automatically to show the new page state.
Research products, book appointments, fill out paperwork, monitor prices, or automate any web task—all through simple text files that anyone can understand.
Star Growth
Repurpose is a Pro feature
Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.
Unlock RepurposeSimilar repos coming soon.