Offprint

Appearance

Capture and verify one page ​

Save a page, open the HTML file, and inspect the evidence that travels with it. Install the CLI before starting, or run the commands with uvx offprint or npx offprint.

Offprint needs network access to the page. It uses a compatible local browser or downloads and verifies the pinned Chrome for Testing build on first use. Chrome for Testing is Google's versioned Chromium browser distribution.

Capture the page ​

console
offprint capture https://example.com --output example.html --quiet

The command prints the saved path:

text
example.html

Open example.html in your browser. The file contains the rendered page and its rendering resources. Offprint already reopened it with network access denied before committing the file. Offline verification is the default.

A capture runs the source page's code inside an isolated browser context. The saved HTML retains captured content and state, while captured page scripts are stripped. Captured private content remains private data.

Repeat a capture ​

By default, an existing destination produces an error. Replace it after the new capture verifies successfully:

console
offprint capture https://example.com \
  --output example.html \
  --on-exists replace

Use --on-exists uniquify to keep both captures under distinct file names.

Read the saved evidence ​

console
offprint artifact inspect example.html

Inspection reads the embedded artifact manifest: source, capture time, browser, resource counts, warning codes, and requested verification mode. It validates the manifest. To check the complete file again, run:

console
offprint artifact verify example.html

Verification checks the HTML structure and embedded resources, then reopens the file in a fresh browser context with network access denied. A successful run reports zero observed network requests.

A self-contained artifact can still have missing content. The default resource policy records warnings and substitutes inert fallbacks when an image, font, or other resource cannot be captured. Use --missing-resources fail when a missing resource must fail the capture.

Continue ​