What goflag is

goflag crawls a site you name by URL and reports four classes of defect:

Broken links: every link on every crawled page, probed and triaged.
Missing translation pages: a route that exists in one locale and not in another, plus the hreflang declarations that are supposed to tie them together.
A robots.txt that contradicts the pages it serves: a site that forbids crawling while every page asks to be indexed.
Missing or misconfigured SEO metadata: a battery of checks on the <head> of every page, each one listed in the rule catalogue.

It needs a URL and nothing else. No account, no API key, no configuration file, no access to your source code.

What it is for

Every defect above is invisible in a browser. The page renders. The layout is right. The defect is in the part of the document nobody looks at, and the cost is paid weeks later in a search index you cannot inspect.

A human can check the <head> of one page. Nobody checks four hundred. That is the whole job.

What it is not

goflag has no dashboard, no hosted service, and no score out of 100. It sends nothing anywhere: the only network traffic it generates is requests to the site you named. Findings name the rule that produced them, so a report can be argued with instead of chased, and the list of what goflag deliberately does not check is in Limits.

The report is the product

The terminal output is a rendering of a JSON document, not the other way round.

goflag https://example.com --json > report.json

That document has a stable shape (see the report reference), and everything else in the tool is built on it: the summary view, the diff against a baseline, the exit code. If you want to feed findings to something other than a person, read the JSON and ignore the terminal entirely.

Where to go next

Install: one command, or a pinned dev dependency.
Quickstart: from a first red run to a gate you trust, in five steps.
Rule catalogue: every rule, what it checks, and why it matters.
CLI reference: every flag, its default, and what it changes.