DMPN: Developer Monitoring & Productivity Nexus

Find your Developer Aura - Compete to be the Sigma Dev

Comment 10

Inspiration

We’ve all had those moments while coding where we think we’re being productive… but somehow two hours pass and all we’ve done was renamed a few variables and watched 50 Instagram reels. We wanted to build a tool that keeps us mindful of our actual productivity, without micromanaging. Great focus and great code quality makes great progress in software development. ⏭️⏭️⏭️

That’s where DMPN – Developer Monitoring & Productivity Nexus – was born: an aura tracker that provides live feedback on whether you’re in the zone 🔎 or zoning out ✖️. Not only is it an aura tracker, but we also provided 🎮gamification🎮 where we can face off against other programmers in ranked aura battles to show who is the most 10x engineer on the planet!

What it does

The core of DMPN is a fine-tuned model through a webcam image that shows the developer's emotions and, and a screenshot of the current screen of the developer's monitor. These images are analyzed in real time using 🦾 OpenAI's vision models 🦾 to generate productivity scores. From there, DMPN gauges a developer’s tech aura – essentially, how focused and productive they are during a coding session. This score is visualized in two✨ clean UIs ✨ to give users a sense of how they’re doing, gently nudging them back on track if needed.

In addition, we also spearheaded ⚔️ Aura Battles ⚔️, an elo-based online matchmaking tool that pitches you against other aspiring developers, in order to see who is the most sigma developer of them all!

How we built it

Check out the second image for our architecture diagram!

We built DMPN as a monorepo with a modern tech stack:

  • Frontend: Built with Next.js and styled using TailwindCSS. We used WebSockets to push real-time productivity scores to the UI. 🌐
  • Backend: Powered by Flask. It handles image capture, preprocessing, and communication with the OpenAI Vision API. 📸

The system captures a screenshot and webcam photo at regular intervals, sending them to the backend for analysis. We designed prompts for OpenAI’s model to interpret visual indicators of productivity or distraction from both face and screen. Productivity scores are calculated and streamed live to the frontend via sockets.

Our design is proud to be UI-first in order to win the UI prize create the best first-time experience for users. In addition, we also prioritize real-time interactivity with the user, optimizing prompting and backend pipeline to provide low latency.

Challenges we ran into

  • Setting up real-time image captured from both the webcam and screen in a smooth, system-agnostic way. 😎
  • Prompt engineering for OpenAI to interpret productivity meaningfully from facial expressions and screen content. 🎭
  • Handling WebSocket communication and ensuring consistent, lag-free updates.📶
  • Balancing user insights with privacy — making sure we don’t overstep while still delivering value.🔐

Accomplishments that we're proud of

  • A functional MVP with end-to-end integration — from image capture to live productivity scoring.
  • Tuned prompts that produced surprisingly accurate productivity assessments. Embrace the exponents!
  • A clean, minimalistic UI that feels futuristic and responsive.
  • Successfully used a monorepo architecture, streamlining development across multiple parts of the stack.

What we learned

  • Vision AI can be surprisingly insightful🧐, but it needs thoughtful prompting and tuning.
  • Real-time interactivity is powerful💪, especially when paired with WebSockets.
  • Modular development and clear collaboration🧑‍🤝‍🧑 made working across the frontend and backend seamless.
  • It's possible to create meaningful productivity tools without being invasive💢 — subtle feedback can go a long way.

What's next for DMPN: My Tech Aura

  • Refined scoring: Add more nuanced categories like deep focus, distraction, passive browsing, etc.
  • Personalization: Let users train a model on their own habits for more tailored feedback.
  • Integrations: Build browser extensions and IDE plugins for smoother workflow monitoring.
  • Privacy-first options: Offer local processing modes with no cloud storage, to ensure full data control for users.

Built With

+ 31 more
Share this project:

Updates