Skip to content
quoxcli

Pull a page's metadata

Machine verified A machine ran this on our install and checked the result against the CLI. A test report is linked. CLI 0.3.1, checked 2026-10-15.

What does this page say about itself?

Hand a page to Quox and get its title, description and share-card tags back, checked against the CLI.

What you get: what a page tells search engines and link previews, read from the live page, not from memory.

Ask Quox

Section titled “Ask Quox”

Name the agent and the page:

quox · talking to Quox

Delegate this to the Browser-Q agent and have it use its own browser tools: extract the page metadata of https://quox.ai. What is the Open Graph title?

▤ AOCL │ L1 identity complete · allowed · 1ms

▤ AOCL │ L3 policy gate complete · proceed · 13ms

◐ AGENT │ directory ✔ AGENT │ directory │ complete: true, 71 total, 71 count · 1.7s

◐ DELEGATE │ delegate_to_agent agent_id=browser-q ✔ DELEGATE │ delegate_to_agent agent_id=browser-q │ 3 tools used, 3 tool outcomes · 92.9s

⛓ AEE │ envelope opened · stream_success

The Open Graph title is Quox · Put AI agents to work, and keep proof of everything.

Browser-Q opened the page in its own browser and read the metadata. The address
redirected from  https://quox.ai  to  https://quox.ai/ .

Other metadata it found:

• og:description: "Run your on-call, ship your releases and answer from your own docs
  with AI agents, with every step witnessed on a chain you can check."
• og:type, og:site_name, og:locale: website, Quox, en_US.
• og:image:  https://quox.ai/og-image.png?v=2 , at 1200×630.
• Twitter card: summary_large_image. Its title, description and image match the Open
  Graph values.
• Canonical URL:  https://quox.ai/ .

The page also carries JSON-LD structured data (Organization, WebSite,
SoftwareApplication and Dataset). Browser-Q can break that out if you want it. The
extraction was read-only and changed nothing on the site.

109.8s · 2 tools · 0 approvals

A real turn on our own install (CLI 0.3.0, 15 Oct 2026), replayed at speed. Names changed. Type the same words into quox.
✔ Machine verified against the CLI on 15 Oct 2026 · test report mv93ae

Quox found the Browser-Q agent and handed it the job. Browser-Q opened the page in its own governed browser, read the metadata, and reported back. The page was only read, never changed.

Do it yourself

Section titled “Do it yourself”

1. Read one tag

Section titled “1. Read one tag”

The attribute inspector reads one attribute of the first element that matches:

og:title 
quox qlarity read https://quox.ai --inspector attribute --selector 'meta[property="og:title"]' --name content🔍 QLARITY: https://quox.ai/ (200) Quox · Put AI agents to work, and keep proof of everything attribute:"Quox · Put AI agents to work, and keep proof of everything"
From a real run on our own install, against quox.ai.

2. Read the others

Section titled “2. Read the others”

Change the selector for the other tags: meta[property="og:description"], meta[property="og:image"], meta[name="twitter:card"], or link[rel="canonical"] with --name href.

What just happened

Section titled “What just happened”

Browser-Q is an agent with its own browser tools. Handing it the job meant the page was opened in an isolated browser on your Quox, and only sites your organisation allows can load there.

Check it yourself

Section titled “Check it yourself”

quox qlarity origins list shows which sites your organisation allows.

Undo

Section titled “Undo”

Reading a page changes nothing, so there is nothing to undo.

Next

Section titled “Next”