Termly

AI-powered covenant monitoring that cuts loan compliance time from 8 hours to 5 minutes.

Comment

Termly

AI-powered covenant monitoring that transforms hours of manual work into minutes.

Category: Keeping Loans on Track

The Problem

Every quarter, credit teams manually verify borrower compliance. The process:

Step Task Pain Point
1 Collect financial statements Chasing borrowers
2 Find covenant definitions in 100+ page documents Time-consuming
3 Extract financial data by hand Error-prone
4 Calculate ratios in spreadsheets Formula mistakes
5 Compare against thresholds Inconsistent interpretation
6 Generate reports Delayed insights

Current process: 5-11 hours per loan, every quarter.

Breaches discovered weeks late. Errors go unnoticed. Teams can't scale.

The Solution

Termly automates covenant monitoring from document to dashboard.

  1. Upload - Drop a credit agreement or compliance certificate (PDF)
  2. Extract - AI reads the document, pulls covenant terms, financials, thresholds
  3. Calculate - System computes ratios and compares against thresholds
  4. Monitor - Dashboard displays compliance status with real-time alerts

Result: ~5 minutes per loan.

Key Features

Document Intelligence

  • Extracts covenant definitions and thresholds
  • Handles EBITDA calculations with permitted add-backs
  • Processes financial data from statements
  • Identifies testing frequency and dates

Automated Compliance Testing

  • Calculates leverage, interest coverage, fixed charge ratios
  • Determines status: Compliant, Warning, or Breach
  • Shows headroom (distance to threshold)

Real-Time Alerts

  • Critical: Covenant breach detected
  • Warning: Headroom below threshold
  • Info: Upcoming test dates

Portfolio Dashboard

Single view of all loans, compliance status, and risk indicators.

Target Users

User Benefit
Credit Analysts Extract data in seconds, not hours
Portfolio Managers See all loans at a glance
Loan Operations Automate routine monitoring
Risk Managers Real-time portfolio visibility

Value Proposition

Before After
5-11 hours/loan ~5 minutes/loan
Weeks to detect breaches Real-time detection
Manual errors Consistent calculations
Spreadsheet chaos Centralized dashboard
More loans = more analysts Scale without adding headcount

Technology Stack

Layer Technology
Frontend Next.js 16, React 19, TypeScript, Tailwind CSS
UI Components Shadcn/ui, Radix UI
Backend Supabase (PostgreSQL), Edge Functions
AI Anthropic Claude, Groq
Document Processing PDF.js, Tesseract.js (OCR)
Auth Clerk
Deployment Vercel

Links
Live Platform termly.cc
Pitch Deck Pitch Deck
Repository github.com/brn-mwai/termly-LMA

Summary
Problem Covenant monitoring is manual, slow, error-prone
Solution AI-powered extraction and real-time monitoring
Result Hours to Minutes
Category Keeping Loans on Track

Built for the LMA EDGE Hackathon 2026

Built With

  • anthropic-claude-api
  • clerk
  • groq
  • next.js-16
  • pdf.js
  • react-19
  • shadcn/ui
  • supabase
  • tailwind-css
  • tesseract.js
  • typescript
Share this project:

Updates

posted an update —

Termly is live

AI-powered covenant extraction and compliance tracking for loan documents.

Built this after studying LMA materials on how covenant monitoring works in practice. Learned that a lot of it still involves manually reviewing agreements and tracking deadlines in spreadsheets.

Tried to build something that simplifies that.

Feedback welcome — especially from anyone in loan operations or agency services. I'm new to this space and want to get it right.

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