livecheck-audit
v0.2.2
Published
Run a fixed technical checklist against any deployed website and get a scored, timestamped report.
Maintainers
Readme
Livecheck
Run a fixed technical checklist against any deployed website and get a scored, timestamped report.
Built for the last mile of AI-built sites: they generate fast, then break at launch. Livecheck finds what broke - DNS, SSL, forms, mobile layout, SEO tags, page speed, security headers - and hands you a shareable report instead of a debugging session.
Quick start
npx livecheck-audit https://yoursite.com LIVECHECK AUDIT
yoursite.com (https://yoursite.com/)
Score: 81/100 - Minor issues before launch
SSL / TLS
PASS Certificate trusted Chain validates against public trust stores
PASS Certificate expiry Expires in 63 days
SEO & Metadata
WARN Meta description Search engines will invent their own snippet
...
12 passed 1 failed 8 warnings 1 skippedExit code is 0 when nothing fails and 1 when at least one check fails, so it drops straight into CI:
- run: npx livecheck-audit https://yoursite.comOptions
| Flag | Description |
| ---- | ----------- |
| --json | Print the report as machine-readable JSON |
| -o, --out <dir> | Save markdown + JSON reports to a directory |
| --probe-forms | Actively submit the first form found (off by default) |
| --ai | Append a plain-language AI analysis of the results (requires an API key, see below) |
| --lighthouse | Add a Lighthouse performance score (requires npm i -D lighthouse chrome-launcher) |
| -t, --timeout <ms> | Per-check timeout in milliseconds (default 20000) |
By default Livecheck is read-only: it never submits forms or mutates anything on the target site. Pass --probe-forms to opt into an active submission test.
What gets checked
| Group | Checks | | ----- | ------ | | General | Site reachable | | DNS & Domain | A/AAAA records resolve, www subdomain resolves | | SSL / TLS | Certificate trusted by public stores, expiry window | | HTTP Headers | HSTS, X-Content-Type-Options, CSP, clickjacking protection, compression | | SEO & Metadata | Title length, meta description, Open Graph tags, single h1, lang attribute, favicon | | Mobile | Viewport meta tag, horizontal overflow at 375px | | Forms | Submission target, submit button present, field labeling, optional active probe | | Speed | Time to first byte, full page load, page weight, optional Lighthouse score |
Scoring
Each check carries a weight (default 1, critical checks 2). Passing earns full weight, warnings earn half, failures earn none, skips are excluded from the denominator. The result is normalized to 0-100:
90+Launch-ready70-89Minor issues before launch50-69Needs work before launch<50Not ready for visitors
AI analysis
Pass --ai to add a plain-language explanation of every failure and warning: what is wrong, why it matters, and how to fix it.
npx livecheck-audit https://yoursite.com --aiThe analysis runs through a pre-configured OpenAI-compatible provider on the machine running the audit (via LIVECHECK_AI_KEY, GROQ_API_KEY, or OPENAI_API_KEY, with optional LIVECHECK_AI_BASE_URL / LIVECHECK_AI_MODEL overrides). No setup is required if you are running reports provided to you.
Only structured check results (titles, statuses, detail strings) are sent to the provider - never page content, credentials, or HTML. Without a key the flag is skipped with a notice and the audit still completes.
Reports
With -o reports/ you get two artifacts per run:
- Markdown - human-readable, timestamped, ready to paste into a ticket or email to a client
- JSON - stable schema for CI gates, dashboards, or your own tooling
Disclaimer
Reports describe observable behavior during a single point-in-time audit. Sites change after they are checked. Livecheck output is provided as-is, without warranty of any kind, and does not certify security or correctness beyond what is listed.
Roadmap
- [ ] Automated pre-intake audit API for service intake flows
- [ ] Auto-fix recipes for the most common deployment failures
- [ ] White-label report exports for agencies
- [ ] Scheduled monitoring with drift alerts
Development
npm install
npx playwright install chromium
npm run dev -- https://example.com
npm test
npm run lint
npm run typecheck
npm run buildA gated end-to-end smoke test against example.com runs with LIVECHECK_SMOKE=1 npm test.
Contributing
See CONTRIBUTING.md. Conventional commits required; npm run verify style checks run in CI.
