AI-Powered Deep QA · Magnus economics: 20% flat, transparent, verifiable
Transparency is the product.
Trust is the moat.
MagnusQA scans any live site with six specialist workers — browser, API, load, security, accessibility, and business-logic. Every bug ships with reproduction steps, screenshots, and a suggested fix. We tell you what we tested, what we found, what it costs, and exactly how the math works.
No card to start. Free tier: 1 site, 3 scans/month, 5-bug preview with honest disclosure of the rest.
Three customers, one engine
Layer 1
Internal ecosystem QA
Every site in the Magnus/Quiky ecosystem is tested by MagnusQA continuously. Internal accounts auto-verify; the rest of the rigor is identical.
Layer 2
Magnus Forge integration
Forge ships sites pre-hardened. Prevention rules from the KB are injected into Forge's build prompt so common bugs never get authored in the first place.
Layer 3
External SaaS
The same engine, available to anyone with a URL. 4 paid tiers, customer-chosen cadence, customer-chosen remediation. No upsell coercion.
How MagnusQA works
Step 1
Onboard in 60 seconds
Enter a URL, accept the Terms of Service, prove you own the domain via DNS TXT, file upload, email, or OAuth. No scan runs before ownership is verified.
Step 2
Pick what matters
Daily, weekly, monthly, or on-demand cadence. Scope to specific paths. Configure auth if you need a logged-in scan. We honor robots.txt and your crawl-rate caps.
Step 3
Get actionable bugs
Every finding includes reproduction steps, before/after screenshots, root cause hypothesis, and a concrete fix. You decide: fix it yourself, dismiss, or hire us (from $50).
The two-tier training loop — the moat
MagnusQA scans your site first. A human external QA firm (Applause, Testlio, QA Wolf) scans afterwards. Any bug they catch that we missed becomes labeled training data for the ecosystem brain. Over the first ~100 sites of the loop, MagnusQA graduates from novice to expert on the bug patterns that matter in real codebases.
Site built → MagnusQA tests first → bugs fixed
│
▼
External QA firm tests same site
│
▼
Bugs external found, MagnusQA missed
│
▼
Training pipeline:
- categorize
- extract pattern
- write detection rule
- write prevention rule
│
▼
Knowledge Base grows, MagnusQA learns,
Magnus Forge ships cleaner.The agent API exports every labeled sample (respecting per-account training opt-out) so the wider ecosystem gets smarter with every scan. Prevention rules are injected into Claude Code and Magnus Forge build prompts so future builders avoid the patterns from day one.
Pricing you can audit
Full pricing →Base tiers priced at sticker rate — the platform is the merchant. Add-on remediation carries the Magnus Economics Charter 20% platform fee. All computationally verifiable at /api/fee-audit/policy.
Free
$0/mo
- · 1 site
- · Surface-level detection
- · 5-bug preview
Starter
$99/mo
- · 3 sites
- · Monthly scans
- · Full bugs + repro
Growth
popular$399/mo
- · 10 sites
- · Bi-weekly cadence
- · Video recordings
Business
$1,499/mo
- · 50 sites
- · Weekly cadence
- · API access
Enterprise
Custom
- · Unlimited
- · Daily scans
- · Dedicated CSM
Transparency commitments
Per the Magnus Manifesto, customers see the whole picture. No upsell coercion, no hidden findings, no surprise fees. These aren’t marketing claims — they’re enforced in the API surface.
- · Every test we ran.
- · Every finding we surfaced (free tier shows a partial set but discloses the rest — never hidden).
- · How severity was scored.
- · Reproduction steps and suggested fixes (paid tiers).
- · The cost if you hire us to fix it.
- · The self-fix approach, always available as an alternative.