🚀 PeerFlow

Split Expenses. Not Friendships.

A graph-optimized peer-to-peer expense settlement engine that uses integer-only paise accounting, a Minimum Cash Flow algorithm (O(n log n)), and democratic governance to minimize transactions and enforce strict financial integrity.

---

🔗 Demo & Resources

Resource	Link
🌐 Live Demo	https://peer-to-peer-expense-splitter-with-chi.vercel.app/
🎥 Demo Video	https://drive.google.com/drive/folders/1yYiBbJ23J51rN05cHb9RtBz5EnbI0yjH?usp=sharing
📦 GitHub Repository	https://github.com/White-Devil2839/Peer-to-Peer-Expense-Splitter-with-Debt-Simplification

---

1️⃣ Problem Statement

Managing shared expenses creates complex debt chains between members.
Most existing tools calculate balances but do not minimize the number of transactions required to settle them.

This results in:

• Redundant transfers
• Circular debt loops
• Floating-point rounding errors
• High coordination effort
• No accountability system

Example

4 members share ₹12,000 across multiple expenses.

Naive settlement: 12 transactions
PeerFlow optimized settlement: 3 transactions

➡ 75% transaction reduction

---

2️⃣ What Makes PeerFlow Different

• Global debt optimization (not pairwise settlement)
• Integer paise accounting (₹100.50 → 10050)
• Minimum Cash Flow algorithm
• Overpayment prevention at both frontend and backend
• 50% expense approval governance
• 75% overdue consensus mechanism
• Σ(net balances) = 0 invariant enforced
• D3.js before/after debt graph visualization
• 56 automated Jest tests ensuring correctness

---

3️⃣ Financial Integrity Model

Integer Accounting

All monetary values are stored as integers in paise.

No floating-point arithmetic is used in financial calculations.

---

Overpayment Protection

The maximum payable amount is computed as:

maxAllowed = Math.min(Math.abs(debtorNet), creditorNet)

Validation rules:

• amount ≥ 1
• amount ≤ maxAllowed
• debtorNet < 0
• creditorNet > 0

If any condition fails, the backend rejects the request.

Frontend protection:

• Input max value dynamically set
• Inline error shown when exceeded
• Simulate button disabled if invalid

Overpayment is mathematically impossible in the system.

---

4️⃣ Minimum Cash Flow Algorithm

Complexity: O(n log n)

Algorithm Steps:

1. Compute net balance for each member
2. Partition into creditors (+) and debtors (-)
3. Sort creditors descending by balance
4. Sort debtors descending by absolute value
5. Match highest creditor with highest debtor
6. Transfer minimum of the two
7. Repeat until all balances become zero

Guarantees:

• Minimum number of transactions
• Deterministic output
• Σ(net balances) = 0 always preserved

---

5️⃣ System Architecture

User → React Frontend → Express API → Financial Engine → MongoDB → Response

Tech Stack:

• React 18 + Vite
• Node.js + Express
• MongoDB + Mongoose
• JWT Authentication
• D3.js Visualization
• Jest Testing

---

6️⃣ ER Diagram

erDiagram
USER {
ObjectId _id PK
String name
String email UK
String password
}
GROUP {
ObjectId _id PK
String name
String joinCode UK
String password
Number settlementThreshold
ObjectId createdBy FK
}
EXPENSE {
ObjectId _id PK
ObjectId group FK
String description
Number totalAmount
ObjectId paidBy FK
String status
Number requiredApprovals
Boolean isRecurring
}
SPLIT {
ObjectId user FK
Number shareAmount
}
APPROVAL {
ObjectId user FK
String vote
}
PAYMENT {
ObjectId _id PK
ObjectId group FK
ObjectId from FK
ObjectId to FK
Number amount
}
OVERDUE_VOTE {
ObjectId _id PK
ObjectId group FK
ObjectId targetUser FK
ObjectId voter FK
String vote
}
GROUP_MEMBER_STATUS {
ObjectId _id PK
ObjectId group FK
ObjectId user FK
String status
}
USER ||--o{ GROUP : "member of"
USER ||--o{ EXPENSE : "paid by"
GROUP ||--o{ EXPENSE : "contains"
GROUP ||--o{ PAYMENT : "contains"
EXPENSE ||--o{ SPLIT : "has"
EXPENSE ||--o{ APPROVAL : "has"
USER ||--o{ PAYMENT : "from"
USER ||--o{ PAYMENT : "to"
GROUP ||--o{ OVERDUE_VOTE : "contains"
USER ||--o{ OVERDUE_VOTE : "target-voter"
GROUP ||--o{ GROUP_MEMBER_STATUS : "tracks"
USER ||--o{ GROUP_MEMBER_STATUS : "has status"

Loading

---

7️⃣ Database Models

User

• name
• email
• password (hashed)

Group

• name
• joinCode
• settlementThreshold
• members[]

Expense

• group
• description
• totalAmount (paise)
• paidBy
• splits[]
• status (pending / approved / rejected)
• approvals[]
• requiredApprovals
• isRecurring
• recurrence

Payment

• group
• from
• to
• amount (paise)

OverdueVote

• group
• targetUser
• voter
• vote

GroupMemberStatus

• group
• user
• status (active / overdue)

---

8️⃣ Governance System

Expense approval:

• Requires ≥ 50% majority

Overdue marking:

• Requires ≥ 75% consensus

Overdue users:

• Cannot create expense
• Cannot vote
• CAN repay debt

---

9️⃣ Testing & Validation

56 Jest tests covering:

• Net balance computation
• Split correctness
• Floating-point prevention
• Overpayment rejection
• Settlement integrity
• Algorithm correctness
• Governance enforcement
• Threshold validation

Invariant enforced:

Σ(net balances) = 0

---

🔟 Team Roles & Responsibilities

Member	GitHub	Role	Responsibilities
Divyansh Choudhary	@White-Devil2839	Full-Stack Engineer	Architecture, backend API, financial engine, governance logic, frontend UI
Ansh Baheti	@techcodie	Algorithm Engineer	Minimum Cash Flow algorithm, balance optimization
Tisha Kharade	@ace-tk	Backend & Testing Engineer	Groups module implementation, testing contributions, deployment setup

---

1️⃣1️⃣ End-to-End Workflow

1. User registers → JWT issued
2. Group created → members join
3. Expense added → status: pending
4. ≥ 50% approval → expense becomes active
5. Net balances computed
6. Minimum Cash Flow applied
7. Simplified graph generated
8. Payment recorded
9. Overdue rules enforced

---

1️⃣2️⃣ Impact

• 60–80% reduction in settlement transactions
• Zero currency rounding drift
• Deterministic financial engine
• Graph theory applied to real-world finance
• Transparent governance for shared expenses

---

1️⃣3️⃣ Known Limitations

• Single currency (INR only)
• No real-time WebSocket sync
• No receipt upload
• No native mobile app

---

1️⃣4️⃣ Future Scope & Scalability

Short-Term

• Progressive Web App (PWA) support
• Push notifications for settlement reminders
• Monthly summary PDF export
• UPI integration for direct settlements

Long-Term

• Multi-currency support
• Predictive expense analytics
• Enterprise-scale architecture
• Role-based permissions
• Real-time synchronization via WebSockets

---

1️⃣5️⃣ Conclusion

PeerFlow demonstrates how graph theory, strict financial invariants, and democratic governance can transform shared expense management into a mathematically optimized and socially fair system.

Every rupee accounted.
Every transaction minimized.
Every decision governed.

---

🔑 Demo Accounts

All demo accounts use the same password: demo1234

Name	Email	Password
Aarav Sharma	aarav@demo.com	demo1234
Priya Patel	priya@demo.com	demo1234
Rohan Gupta	rohan@demo.com	demo1234
Sneha Iyer	sneha@demo.com	demo1234
Kabir Mehta	kabir@demo.com	demo1234

---

🛠️ Local Setup

# Clone the repository
git clone https://github.com/White-Devil2839/Peer-to-Peer-Expense-Splitter-with-Debt-Simplification.git
cd Peer-to-Peer-Expense-Splitter-with-Debt-Simplification
# Backendcd server
npm install
cp .env.example .env # configure MONGODB_URI, JWT_SECRET, CLIENT_URL
npm run seed # seed demo data
npm run dev # starts on port 5005# Frontend (new terminal)cd client
npm install
npm run dev # starts on port 5173

---

Built for OverClock 2026
Graph Theory × Financial Integrity