Inspiration 💡
It started with a $25,000 "administrative error."
A friend working at a promising HealthTech startup once shared a nightmare: they missed a single HIPAA re-certification deadline because the notification email was buried in a "Promotions" tab. No one was "watching" the watchers. That one missed date led to a $25,000 fine and weeks of legal audits.
We realized that in the world of high-stakes compliance—GDPR, SOC 2, HIPAA—the current tools are either boring spreadsheets that everyone ignores or passive dashboards that no one checks.
Vigil was born from a simple question: What if your compliance system wasn't just a list, but a persistent digital butler that guarded your business like a hawk?
What it does 🔍
Vigil is an AI-powered Compliance Agent built for the modern, high-speed enterprise.
Imagine you've just received a 50-page PDF audit or a complex URL full of regulatory requirements. Instead of manually mapping deadlines into a calendar:
Drop & Ingest: You feed the document to Vigil. AI Extraction: Using Google Gemini, Vigil "reads" the legalese, identifies every mandatory requirement, and extracts the exact deadlines. The Vigilance Score: Your company is assigned a real-time health score. If you ignore a critical GDPR update, your UI turns from a calm "Electric Blue" to a "Critical Red," and the Vigilance Score drops. Active Nagging: Vigil doesn't just notify; it escalates. It acts as an "Active Agent" that ensures requirements are met before they become liabilities.
How we built it 🛠️
We leaned heavily into the Serverpod ecosystem to build a truly full-stack Dart application.
The Brain: We integrated Google Gemini Pro directly into our Serverpod endpoints. This allows Vigil to perform deep semantic analysis on any document or URL provided by the user. The Heart: Serverpod manages our database and type-safe communication. Defining a "Requirement" in the backend meant it was instantly available in our Flutter frontend with no boilerplate. The Shield: The Flutter frontend is a premium, glassmorphic command center. We used custom painters for the Vigilance Score Gauge and implemented smooth, high-fidelity transitions to make "Regulatory Work" feel like "Mission Control."
Challenges we ran into 🚧
The biggest hurdle was Semantic Accuracy. Compliance isn't just about finding dates; it's about understanding context. We spent hours perfecting our AI prompts to ensure that when Vigil sees "shall be completed within 30 days of the fiscal year-end," it accurately calculates the date based on real-world time.
We also faced the "Hackathon Deadline" challenge—ensuring that if a judge tests the app while our local server is sleeping, the experience doesn't break. We implemented a robust "Demo Mode" that seamlessly swaps in realistic compliance scenarios (GDPR, SOC 2) if the backend is unreachable.
Accomplishments that we're proud of 🏆
We successfully built a Type-Safe, AI-Driven Agent in under 48 hours. The way the "Vigilance Score" dynamically updates and changes the entire aesthetic of the app is something we're incredibly proud of. It turns "boring compliance" into an engaging, high-stakes experience.
What we learned 🎓
We learned that Serverpod is a superpower. The speed at which we could iterate on backend models and have them reflected in our Flutter web build allowed us to spend 80% of our time on "Cool Features" and only 20% on "Wiring."
What's next for Vigil 🚀
Vigil is just getting started. Our roadmap includes:
Auto-Remediation: Vigil drafting the actual compliance reports for you to review. Agent Escalation: If a deadline is 2 hours away, Vigil will "level up" its nagging—calling phones or locking Slack channels until the requirement is satisfied. Multi-Regime Support: Specializing the Butler for specific laws like the EU AI Act or CCPA. Vigil: Because in compliance, the only safe score is 100%.
Built With
- dart
- docker
- flutter
- google-cloud
- google-gemini
- postgresql
- serverpod


Log in or sign up for Devpost to join the conversation.