Understanding Crawl Results
Read scan results: pages, states, accessibility, SEO, security, performance, AI test coverage banner, baselines, and comparisons.
Understanding Scan Results
When a scan finishes, the result page is your one-stop view of everything AegisRunner found: pages, the things on each page worth testing, audit findings, and the AI-generated test suites. This guide walks through every section.
The header
The top of the page summarizes the run:
- Pages — distinct URLs the scanner reached.
- States — distinct interactive states across those pages (a page with a modal is two states; a page with three tabs is three).
- Duration — wall-clock time end to end.
- Coverage banner — "AI tests cover X of Y discovered pages." If you're on the free plan or a thin site, this tells you exactly which pages got tests and which didn't, with an upgrade link if it's plan-bound.
- Set as Baseline / Compare with Baseline — once you have a clean scan, lock it in as your project's baseline so future scans can be compared and CI can replay deterministically.
The tabs
| Tab | What's in it |
|---|---|
| Pages | Every page found, with its screenshot, state count, and status. Click a row to drill in. |
| Accessibility | WCAG findings rolled up across the site. Severity, affected pages, and fix guidance. |
| SEO | Per-page meta tags, heading structure, missing alt text, structured-data warnings. |
| Security | Security headers across the site (HSTS, CSP, X-Frame-Options, Referrer-Policy, etc.). |
| Performance | Load times and resource sizes per page, sorted slowest first. |
| Forms | Every form the scanner detected, with field counts and a classification (login, search, contact, checkout, etc.). |
| Fill Results | If you ran with Fill Forms enabled, what got filled, what got submitted, and what each submission produced. |
| Artifacts | Raw outputs — sitemap, link graph, console logs, network captures. |
| Test Suites | The AI-generated suites for this scan, with the coverage summary repeated up top. |
| Dead Elements | Buttons or links that did nothing when clicked — common signal of broken UI or unhandled error states. |
| Recommendations | Page-level recommendations across all tabs, prioritized. |
Pages and states
Most pages have more than one state. The first state is the page as it loads. Each interaction (clicking a button, opening a dropdown, filling a form) can produce a new state.
| State type | Example |
|---|---|
| Initial | The page as it first loads. |
| Update | Content changed on the same page after an interaction. |
| Modal | A modal or dialog opened. |
| Expand | An accordion, tab panel, or details element expanded. |
| Form validation | Validation errors appeared after a submit. |
| Navigate | An interaction took the user to a different page. |
Page detail
Click any page from the Pages tab to see:
- Screenshots for each state, with the action that triggered it.
- Interactive elements — every button, link, and input the scanner found, with the locator it'll use in tests.
- Page-level audits — accessibility, SEO, security, performance findings scoped to this page.
- Page recommendations panel — actionable items, prioritized.
- Test suites for this page — every suite that was generated from it, with a button to open the suite.
The AI coverage banner
On the result page header and on the Test Suites tab, you'll see something like:
"AI tests cover 4 of 12 discovered pages."
This exists because the relationship between "pages found" and "tests generated" isn't always 1:1. A few reasons coverage might be partial:
- Free plan — only the entry page gets AI tests. Free scans always show the upgrade CTA next to the banner.
- Pages with no actionable elements — a page that's pure decorative content has nothing meaningful to test, so the AI skips it. (Free-tier scans of thin pages still get a "smoke test" filler — page loads, mobile viewport, top link works — so you always have something.)
- AI generation in progress — for large scans, suite generation continues after the scan completes. Coverage rises over the next minute or two.
Setting a baseline
Once a scan looks clean — the right pages, the right interactions, no obvious flakes — set it as the baseline:
- Click Set as Baseline at the top of the result page.
- AegisRunner compiles a manifest of pages, interactions, and expected screenshots.
- Baseline Replay mode now appears in the scan modes for this project — use it from CI for deterministic checks.
- Compare with Baseline appears on every future scan.
Comparing against the baseline
On any non-baseline scan, click Compare with Baseline to see drift:
- Missing pages — were in the baseline, gone in this scan. Often a regression.
- New pages — appeared in this scan but weren't in the baseline. Often expected for a release.
- Changed pages — pages with different state counts, different interactions, or different screenshots.
- Unchanged — matched the baseline exactly.
From scan results to running tests
The AI-generated suites land in Test Suites in the sidebar (or the Test Suites tab on the scan result). From there:
- Running Tests — pick browsers, run, watch results stream live.
- Test Baseline Review — sanity-check AI tests before you trust them.
- Visual Regression — turn on pixel-level checks for the next run.