sigilbox/README.md

# Local Page Archiver

This project saves self-contained HTML archives for pages the operator is authorized to access. It sends a real browser user agent, renders web URLs with Playwright, strips ad/tracker-like elements, normalizes the captured DOM, and inlines page requisites as `data:` URLs.

It intentionally does not execute paywall-bypass rules. The bundled `bypass-paywalls-clean-filters` files are treated as reference material only; paywall selectors and scripts are not applied.

## CLI

```sh
npm install
npm run install-browsers
node src/cli.mjs archive "https://example.com/article"
```

For an existing HTML file:

```sh
node src/cli.mjs archive ./page.html --static
```

For an `archive.ph` HTML export where you want the captured page without the archive shell:

```sh
node src/cli.mjs archive ./bloomberg-archive.html --static --strip-archive-shell
```

Local `archive.ph` HTML inputs with `--strip-archive-shell` use the static extractor by default because those files already contain the rendered page. Add `--render` only when you explicitly want Chromium to load the local HTML first.

Computed-style freezing is off by default for live web pages because it can inflate modern article pages into very large HTML files. Add `--freeze-styles` only when stylesheet inlining is not enough to preserve layout.

Archives are written to `ARCHIVE_PATH`, or to a development directory under the system temp directory when `ARCHIVE_PATH` is not set.

## API

```sh
ARCHIVE_PATH=/tmp/local-page-archives npm run serve
```

Archive a page:

```sh
curl -X POST http://127.0.0.1:8787/archive \
  -H 'content-type: application/json' \
  -d '{"url":"https://example.com/article"}'
```

The response includes the archived file path and a local `viewUrl`.

Set `PORT` to choose a port other than the default `8787`.