Web archiver · capture for agents

Capture the page.
Keep it whole.

A dependency-free Chrome extension that saves any web page as a faithful archive — for humans to revisit and for AI agents to read.

View on GitHub ↗ Install

● LIVE 1.2 MB · 42 assets · zh-CN · UTF-8

FIG. 01 — capturing

def hello(): return "snap"

a page, framed and scanned

Three formats

.MHTMLMost faithful. Chrome's native capture — exact page state, browser-resolved resources.mhtml

.SINGLESelf-contained. One .html with every resource inlined as data URLs.html

.ZIPPortable. index.html + text + messages + resources, ready for agents.zip

Every archive carries an agent-readable layer. Code blocks keep their indentation; conversation pages come out as structured turns. Real specimens:

text-content.txt

What is 1 + 1?

It is 2.

def add(a, b):
    return a + b

Non-destructive · code intact

messages.json

[
  { "role": "user",
    "text": "What is 1+1?" },
  { "role": "assistant",
    "text": "It is 2." }
]

DOM-based turns · ChatGPT & Claude

Then open any page, hit the PageSnap icon, pick a format, save. Full details in the README.