How the risk score works
The tool runs your pasted contract through about 35 pattern checks. Each match contributes to a category score: termination, IP, payment, liability, confidentiality, non-compete, scope, and a few miscellaneous flags. Positive signals — explicit kill fees, mutual indemnification, capped liability, defined dispute resolution — subtract from the risk total.
The final 0–10 score is the weighted sum, normalized so that 0–3 = generally balanced, 4–6 = lean against you, negotiate before signing, and 7–10 = heavily one-sided, don't sign without serious revisions. The category breakdown tells you where the risk concentrates, which is what you actually need to negotiate.
What this tool does — and doesn't — catch
Catches reliably: auto-renewal traps, missing kill fees in freelance contracts, IP work-for-hire over-reach, Net-60+ payment terms, perpetual confidentiality, broad non-competes, mandatory arbitration with class-action waivers, indemnification asymmetry, missing liability caps, undefined scope, vague termination conditions.
Doesn't catch: nuanced "reasonable" language that's unreasonable for your specific deal, jurisdiction-specific issues, regulatory compliance gaps (HIPAA, GDPR, financial services), contract math errors, schedule and exhibit consistency, implication chains across multiple referenced documents.
For the things this tool doesn't catch, use Scrutr's full AI review — it reads the contract holistically and surfaces issues a regex can't see. For high-stakes deals (anything over $50K of contract value, or fundraising/M&A/IP licensing), use a lawyer.
Privacy: how this stays private
The tool is pure client-side JavaScript. The contract text you paste lives in your browser tab and nowhere else. Scrutr's servers don't see it; we don't log it; we don't store it. You can verify by opening your browser's DevTools → Network panel and watching: no network requests fire when you click "Score." The only requests on this page are the initial page load and font/CSS resources.
If you want a more thorough AI-powered review, that is sent to our backend (and from there to Anthropic's API). That's a separate product. Click "Get the full review" only when you've decided you want it.
What contract types does this handle?
The pattern library is tuned for English-language US contracts but most flag categories apply internationally. Best results: SaaS subscriptions, master service agreements (MSAs), freelance and independent contractor agreements, NDAs (mutual and one-way), employment offer letters, residential leases, commercial leases, software license agreements, and consulting agreements. Specialty contracts (M&A, IP licensing, financial derivatives, regulated industry agreements) need lawyer review regardless of what this tool says.
Common questions
Is the score deterministic?
Yes. The same contract pasted twice produces the same score every time. There's no randomness or AI-generated variance. The patterns are static rules; if you find a clause the tool incorrectly flagged or missed, it'll do the same on every contract with similar language.
Can I use this on confidential contracts?
Yes — that's the design. Because it's fully client-side, you can paste confidential contracts, NDAs you've received, or contracts under existing confidentiality obligations without leaking anything. If you're truly paranoid, run it offline by saving the page locally first.
Why is the AI version paid?
The full AI review uses Claude (an LLM that costs real money per request) and produces a much deeper analysis: clause-by-clause findings, custom negotiation email drafted in your voice, inline redlines, and shareable report links. Free tier covers one contract per month. Pro is $29/mo for 100 contracts. See pricing.