// Case study details
Monytrace
A hackathon project to track token usage across various AI platforms, detecting token spend anomalies, and sending automated alerts to prevent budget overruns.
Created on
October 2025
Industry
Finance
Category
Vibe coding
Platform
Tech stack
OpenAI + Lovable
Have a question?
Our friendly team is always here to help you answers whenever needed.
The Challenge
In October we participanted in a Fintech hackathon in San Francisco. Just weeks before the event, one of our projects had seen OpenAI costs spike 154% in a single day due to a runaway background process. By the time we noticed, we had burned through three days of our monthly budget in 6 hours. The bill was painful.
We weren't alone. Every team using multiple AI services faces the same challenge: unpredictable costs, scattered billing systems, and reactive cost management. Our idea was simple: build a unified dashboard that monitors all AI spending in real-time, detects anomalies immediately, and alerts teams before budget overruns happen.
Our Solution
We had 24 hours to prove we could solve this problem. Using Lovable's AI-powered development environment, we built MoneyTrace from concept to working demo.
We created a unified dashboard connecting to OpenAI APIs to aggregate real-time cost data. We implemented anomaly detection that triggers alerts when daily spend exceeds 2x the 3-day average. We integrated Slack notifications delivering contextual alerts directly to team channels with specific issues, potential causes, and recommended actions. Teams get immediate warnings when costs spike, not days later. We added automated cost-saving recommendations including identifying idle EC2 instances ($5.2/hour saved) and suggesting OpenAI model downgrades. Each action is one-click to implement. Finally, we ran MoneyTrace through NEKOD's governance platform to ensure security and production readiness, fixing critical issues including API key management and error handling.
The Results
1st 🏆
Hackathon ranking
24 hours
Build time
∞
Love received
Outcomes and next steps
MoneyTrace won first place and validated strong market demand for AI cost intelligence. Users can save AI tokens through automated alerts, catching runaway processes within hours instead of at month-end. Teams can see total AI spend across all vendors in one place, with budget runway visibility showing days remaining at current rate.
Key Learnings
Hackathons are fun again! What might take weeks in traditional development was accomplished in hours with vibe coding.
Running Monytrace through NEKOD revealed hardcoded API keys, incomplete error handling, and missing monitoring. Fixing these prevents security issues and showed production-grade thinking.
Every team we talked to faces unpredictable AI costs, IRL this problems remains to be solved. Monytrace was discontinued after the hackathon. If you want to make it yours, get in touch.
// Case study details
// Case study details
Monytrace
A hackathon project to track token usage across various AI platforms, detecting token spend anomalies, and sending automated alerts to prevent budget overruns.
Category
Vibe coding
Platform
Created on
October 2025
Tech stack
October 2025
Have a question?
Our friendly team is always here to help you answers whenever needed.
The Challenge
In October we participanted in a Fintech hackathon in San Francisco. Just weeks before the event, one of our projects had seen OpenAI costs spike 154% in a single day due to a runaway background process. By the time we noticed, we had burned through three days of our monthly budget in 6 hours. The bill was painful.
We weren't alone. Every team using multiple AI services faces the same challenge: unpredictable costs, scattered billing systems, and reactive cost management. Our idea was simple: build a unified dashboard that monitors all AI spending in real-time, detects anomalies immediately, and alerts teams before budget overruns happen.
Our Solution
In October we participanted in a Fintech hackathon in San Francisco. Just weeks before the event, one of our projects had seen OpenAI costs spike 154% in a single day due to a runaway background process. By the time we noticed, we had burned through three days of our monthly budget in 6 hours. The bill was painful.
We weren't alone. Every team using multiple AI services faces the same challenge: unpredictable costs, scattered billing systems, and reactive cost management. Our idea was simple: build a unified dashboard that monitors all AI spending in real-time, detects anomalies immediately, and alerts teams before budget overruns happen.
The Results
1st 🏆
Estimated savings
24 hours
Delivery time
∞
ROI
Lessons and outcomes
Hackathons are fun again! What might take weeks in traditional development was accomplished in hours with vibe coding.
Running Monytrace through NEKOD revealed hardcoded API keys, incomplete error handling, and missing monitoring. Fixing these prevents security issues and showed production-grade thinking.
Every team we talked to faces unpredictable AI costs, IRL this problems remains to be solved. Monytrace was discontinued after the hackathon. If you want to make it yours, get in touch.
// Case study details
// Case study details
Monytrace
A hackathon project to track token usage across various AI platforms, detecting token spend anomalies, and sending automated alerts to prevent budget overruns.
Category
Vibe coding
Platform
Created on
October 2025
Tech stack
October 2025
Have a question?
Our friendly team is always here to help you answers whenever needed.
The Challenge
In October we participanted in a Fintech hackathon in San Francisco. Just weeks before the event, one of our projects had seen OpenAI costs spike 154% in a single day due to a runaway background process. By the time we noticed, we had burned through three days of our monthly budget in 6 hours. The bill was painful.
We weren't alone. Every team using multiple AI services faces the same challenge: unpredictable costs, scattered billing systems, and reactive cost management. Our idea was simple: build a unified dashboard that monitors all AI spending in real-time, detects anomalies immediately, and alerts teams before budget overruns happen.
Our Solution
In October we participanted in a Fintech hackathon in San Francisco. Just weeks before the event, one of our projects had seen OpenAI costs spike 154% in a single day due to a runaway background process. By the time we noticed, we had burned through three days of our monthly budget in 6 hours. The bill was painful.
We weren't alone. Every team using multiple AI services faces the same challenge: unpredictable costs, scattered billing systems, and reactive cost management. Our idea was simple: build a unified dashboard that monitors all AI spending in real-time, detects anomalies immediately, and alerts teams before budget overruns happen.
The Results
1st 🏆
Estimated savings
24 hours
Delivery time
∞
ROI
Lessons and outcomes
Hackathons are fun again! What might take weeks in traditional development was accomplished in hours with vibe coding.
Running Monytrace through NEKOD revealed hardcoded API keys, incomplete error handling, and missing monitoring. Fixing these prevents security issues and showed production-grade thinking.
Every team we talked to faces unpredictable AI costs, IRL this problems remains to be solved. Monytrace was discontinued after the hackathon. If you want to make it yours, get in touch.
For organisations building apps faster with AI.
Partner with NEKOD to turn your vibe coded apps into production grade solutions that drive measurable results.
// Get started
For organisations building apps faster with AI.
Partner with NEKOD to turn your vibe coded apps into production-grade solutions that drive measurable results.
// Get started
For organisations building apps faster with AI.
Partner with NEKOD to turn your vibe coded apps into production-grade solutions that drive measurable results.