How Veriify works
A guided tour of the whole platform: what each page does, how the diagnose-and-fix loop works, and where to go next.
Scan, score, fix, verify
Every Veriify workflow, free or paid, moves through the same four steps.
Scan
Point Veriify at a URL or connect a GitHub repo. It crawls every page, form and asset and runs ${WEB_DIMS} quality & integrity dimensions in parallel.
Score
Get one Product Health Score (0-100) with a full dimension breakdown. Every deduction traces to a specific, reproducible finding with evidence. Never a black box.
Fix
Open the Fix Workbench to see every finding with a proposed remediation. Edit code yourself in the Scoped Editor, get the AI Agent's help, or hire a vetted specialist.
Verify
Nothing is marked fixed on your word. Veriify re-scans and only a passing check, or a live sandbox verification, moves your score up and releases any escrow payment.
A fast, static crawl
The free scan is instant and needs no signup for your first result. Veriify crawls every page it can find, checks links, forms, headers, meta tags and more, and scores what a static pass can see across all 9 dimensions.
Runs your product in a real browser
Deep Dive boots an actual headless browser and loads your product like a real visitor would, catching what a static crawl structurally cannot: JavaScript errors, Content-Security-Policy blocks, broken renders, Core Web Vitals (LCP and layout shift), low-contrast text, and other runtime-only bugs. Start it from the Deep Dive button next to the scan box on the Scan Room or the homepage. It takes a bit longer since it's really running your app, and results open straight into the Fix Workbench.
More than "does it work", is it honest?
Veriify scores integrity alongside quality, checks any site publicly, and tests AI products, not just websites.
Genuineness & Integrity
A ninth dimension scores honesty: dark patterns, fake urgency, hidden fees, real-business signals, and whether AI use is disclosed.
Public trust check
Anyone can check any site at /check: HTTPS, security, and trust signals, stated as facts. Owners earn a verified badge and a Registry listing.
AI-product testing
Point Veriify at a chatbot or agent (API or chat page). It probes for jailbreaks, unsafe output, inconsistency and, with your docs, accuracy, for one AI Trustworthiness score with every finding backed by its transcript.
AI transparency
Declare how your product uses AI; Veriify audits the declaration against its scans and publishes the result on your public check page.
Where everything lives
A first-time map of the product, click any card to jump straight there.
Scan Room
Run a free scan or Deep Dive, and pick up recent results. This is where you land once signed in.
AI-Product Test
Behavioural testing for chatbots and agents: jailbreaks, safety, consistency, accuracy.
Trust Check
Check any website's public trust signals, and see your own verified badge.
Registry
Browse platforms that have verified their identity and quality with Veriify.
Documentation
The complete platform reference: scoring, every check, scanning, fixing, and building with Veriify.
Dimensions
The 9 scoring dimensions and how weights adapt by product category.
Rubric
The full, versioned list of every check Veriify runs and how each one is weighted.
Marketplace
Browse vetted specialists and hire one to fix findings the Agent can't handle alone.
Pricing
Compare Free, Pro, Premium and Done-for-you, including the 7-day Premium trial.
Account & scans
Your profile, billing, verified domains, and the full history of every scan you've run.
Notifications
Updates on engagements, fixes, and re-scans that affect your score.
Where every finding gets fixed
Every finding on your scan report lives in the Fix Workbench (Pro and up) with the evidence, a proposed fix, and a risk level. This is the one page where diagnosis turns into action.
- Open a finding to see exactly what's wrong and why it matters.
- Click Edit in sandbox to open the exact flagged file in the in-browser Scoped Editor. Never the rest of your codebase.
- Ask the Agent chat for help, or write the fix yourself.
- Verify on server re-checks your edit against the same rule that flagged it.
- Deploy opens a real pull request on GitHub, reviewed, never force-merged.
When you don't want to fix it yourself
Premium's autonomous Agent triages every finding: it fixes and verifies what it safely can on its own, and escalates the rest with a suggested specialist.
- Hiring creates a pending engagement, no money moves yet.
- The specialist reviews the brief and either accepts (after signing an NDA) or declines.
- Once accepted, you fund escrow, the charge is authorized, not released.
- The specialist works in a scoped sandbox: only the isolated problem, never your full codebase.
- A passing re-scan releases payment automatically. You never pay for an unverified fix.
Common questions
Do I need an account to run a scan?
No, the first free scan works without signing up. Create an account to save history, unlock every finding, and use the Fix Workbench.
What's the difference between the free scan and Deep Dive?
The free scan is a fast, static crawl: and it already checks for known-vulnerable JS libraries, email-security (SPF/DMARC), TLS-certificate expiry and more. Deep Dive (Premium) additionally runs your product in a real browser, catching runtime issues (JS errors, CSP blocks, broken renders, Core Web Vitals and low-contrast text) a static scan can't see.
Can I check a site I don't own?
Yes: veriify.io/check/anydomain.com shows any site's public, observable signals (HTTPS, security headers, a discoverable privacy policy, and more) stated as facts, never a verdict. Our browser extension shows the same check for whatever site you're on. Sites whose owners verify ownership earn a badge and appear in the Registry.
Can Veriify test an AI chatbot or agent, not just a website?
Yes. AI-product testing converses with your product (via its API or chat page) and probes for jailbreaks, unsafe output, inconsistency, and, if you supply docs or Q&A, accuracy, producing one AI Trustworthiness score with every finding backed by its transcript.
Can I just ask Veriify to do things?
Once you're signed in, the Veriify Agent is on every page (bottom-right). Ask it to scan a site, explain any finding or check, or point you to the right tool. It can run a scan for you and link straight to the report.
Is my code ever exposed to a specialist?
No. Specialists work in a scoped sandbox seeing only the single flagged file, never your repository. They also sign an NDA before gaining any access.
Where do I see my score history or re-run a scan?
Account & scans lists every scan you've run with its score and date; open any of them to re-scan in place.
How do I get to this guide again?
"How it works" is always in the top navigation on every page.