Context

Most personal finance apps assume users are comfortable with manual data entry, financial terminology, and abstract categorization. In practice, this creates friction and leads to abandonment.

 
Real-world spending is contextual:

• money is paid from somewhere
• to someone or somewhere
• through a familiar routine

 
Mudraa was built to reflect this reality while remaining fully usable without internet connectivity.

---

Problem

The core problem was enabling fast, repeatable expense tracking without requiring users to think in financial abstractions.

 
Key constraints:

• Users do not want to manually type repetitive data
• Expense tracking must remain usable offline
• Insights must stay understandable without complex dashboards
• Performance must remain instant with growing local data

 
Most existing apps optimized for features rather than behavior.

---

Approach & System Design

The application is built on a local-first, entity-driven model.

 
Instead of flat expense records, transactions are structured around three primary entities:

• Accounts: cash, bank, wallet, loans, or external accounts
• Persons: individuals money is sent to or received from
• Stores: real-world purchase locations with reusable items

 
This structure allows transactions to mirror how users naturally think about spending.

 
The architecture prioritizes:

• local storage as the source of truth
• predictable performance
• minimal user input for repeated actions

 
Stacked (MVVM) was used to enforce separation between UI, business logic, and data handling as the feature set expanded.

---

Transaction Modeling

The system supports three explicit transaction flows:

1. Store-based purchases

   • Select store and predefined items
   • Quantities auto-calculate totals

2. Person-based transfers

   • Send or receive money
   • Manual amount entry with context

3. Account transfers

   • Move funds between internal or external accounts

 
Each transaction supports:

• custom amount handling
• date and time selection
• optional notes for traceability

 
This structure reduces friction while maintaining accurate records.

---

Smart Suggestions

To minimize input effort, Mudraa includes logic-based suggestions, not AI models.

 
Suggestions are generated using:

• frequency of past usage
• recency and time context
• transaction patterns

 
Stores, items, accounts, and people adapt over time, allowing faster entry without sacrificing control or transparency.

---

Insights & Feedback

Insights focus on clarity rather than volume.

 
Key views include:

• income and expense totals
• account balance distribution
• spending health states based on daily limits
• debit and credit trends over time
• top spending categories

 
All visualizations are designed to remain understandable at a glance without financial expertise.

---

Outcomes

• Consistent 5★ rating from active users
• Valued for speed, clarity, and realistic structure
• Fully usable offline with encrypted backups
• Stable performance with large local datasets

 
Mudraa evolved through multiple iterations into a reliable personal finance tool aligned with real-world usage patterns.

---

Trade-offs & Lessons

• Chose structured modeling over flexible free-form entry
• Limited feature expansion to preserve clarity
• Prioritized instant performance over advanced analytics

 

If revisiting early versions, I would formalize migration strategies sooner to simplify long-term schema evolution.