The latest articles on DEV Community by Monsiu Tech (@monsiutech). https://dev.to/monsiutech https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4022997%2Fc30bc8e8-b1f4-4d8a-be6b-64890ebf3d4e.jpg https://dev.to/monsiutech en Monsiu Tech Thu, 09 Jul 2026 15:44:44 +0000 https://dev.to/monsiutech/i-built-an-ai-contract-reviewer-that-has-to-cite-its-sources-heres-the-architecture-40ek https://dev.to/monsiutech/i-built-an-ai-contract-reviewer-that-has-to-cite-its-sources-heres-the-architecture-40ek <p>AI contract summaries share one flaw: you can't tell when they're wrong. The model says "this lease has an aggressive auto-renewal," and you either trust it or read the whole contract anyway, which defeats the point.</p> <p>So I built <a href="https://clauseshift.com" rel="noopener noreferrer">ClauseShift</a> around one rule: <strong>every risk it flags must quote the exact clause it came from, verbatim.</strong> If the model can't point to the wording, it doesn't get to make the claim.</p> <h2> The cross-check architecture </h2> <p>The interesting part is the premium pass: <strong>two models (Gemini 3.1 Pro and DeepSeek V4) review the same contract in parallel</strong>, then a consolidation step merges their findings and computes a combined risk score. Where they disagree, the report says so instead of hiding it. Disagreement between models turns out to be signal, not noise: if only one model flags a clause, that's a "read this yourself" marker.</p> <p>A few hard-won lessons:</p> <p><strong>1. Structured output needs an escape hatch.</strong> Both models return JSON memos (findings, risk enums, key dates). The JSON keys and risk levels stay English even when the report language is German or Japanese; only the prose is localized. Mixing schema and language in one instruction was the #1 source of malformed output.</p> <p><strong>2. Fallback chains fail silently.</strong> My last-resort fallback (direct Gemini Flash) was broken in production for weeks and I only found out by deliberately breaking my primary API key in a test. The bug: Gemini 2.5+ are thinking models, and the thinking tokens consumed the entire output budget, so responses came back empty. One thinkingBudget cap fixed it. Test your fallbacks by breaking things on purpose.</p> <p><strong>3. Fairness beats margins.</strong> If one of the two premium models fails mid-run, the user gets the surviving model's report, a notice that it fell back, and their credit is not charged. That policy cost me an hour to build and is the kind of thing that makes people trust a paid product.</p> <h2> Stack </h2> <ul> <li> <strong>Web</strong>: Next.js (App Router) on Railway, one app serving the marketing site and the product split by hostname in middleware</li> <li> <strong>Mobile</strong>: Flutter, one codebase for Android (Play, Amazon, Huawei) with per-store billing behind build flags</li> <li> <strong>Data</strong>: Supabase (auth + entitlements), reports stored on-device by default with opt-in sync</li> <li> <strong>Payments</strong>: Paystack on web, native billing in each store</li> </ul> <h2> Privacy design </h2> <p>Contracts are the documents people least want in a chatbot log. So: no training on user contracts, uploaded files are not retained after the review, reports live on the device unless you opt into sync, and account deletion actually deletes.</p> <p>The free tier is 3 reviews a month (no card). If you review contracts regularly, I'd genuinely like to know where it falls short: the two-model cross-check either earns your trust or it's just doubled inference cost, and I want to find out which.</p> ai nextjs flutter showdev