Promptivore - Mindful AI usage, visualised

Promptivore helps you track the carbon footprint of your AI use. See emissions, set goals, earn XP. No guilt - just simple, clear feedback.

Comment 1

Inspiration
The idea for Promptivore came from growing awareness around the environmental impact of artificial intelligence. While AI has huge potential, its increasing compute demand also contributes to rising carbon emissions. I saw the need for a simple tool that helps developers and everyday users track their AI usage and understand its environmental cost. The aim was to encourage more mindful AI use—without guilt or overcomplication—and to make the experience engaging through a bit of gamification.

Carbon Offset Calculator

Real-world CO₂ impact equivalents and offset info

What it does
Promptivore is a web app that helps users track, visualise and reduce the carbon footprint of their AI use. It lets you:

  • Track AI usage: Log sessions by model, token count and task type.
  • Calculate CO₂ emissions: Estimate emissions per session using industry data.
  • Real-world comparisons: Translate numbers into something relatable (e.g. charging a phone).
  • Smart model suggestions: Recommend more efficient models based on your usage and task.
  • Token calculator: Estimate token counts across models to optimise prompts.
  • Interactive dashboard: Visualise trends, breakdowns by model and task, and track goals.
  • Session history: Filter, sort and export a log of all past interactions.
  • Gamified experience: Earn XP and level up through efficient, eco-conscious AI use.
  • Eco-badges: Get personalised badges based on your overall emissions.
  • Tips and carbon offset info: Find practical ways to cut impact and explore offset options.
  • User accounts: Manage your data privately and securely.

Gamification System

XP, streaks, and eco-badges for mindful AI use

How I built it
Promptivore is a modern web app built with:

  • Frontend: React + TypeScript for structure and scale.
  • Styling: Tailwind CSS for quick, consistent design.
  • Animations: Framer Motion for smooth transitions and feedback.
  • Icons: Lucide React for clean, customisable icons.
  • Charts: Recharts to render interactive graphs and trends.
  • Backend: Supabase for the database and secure user auth.
  • Calculations: Custom TypeScript functions to estimate emissions and provide equivalents.
  • Persistence: Supabase + localStorage to manage data across sessions.
  • Deployment: Hosted on Netlify for speed and simplicity.

Challenges I ran into
A key challenge was estimating emissions for different AI models. These can vary depending on things like the energy source of the data centre and hardware used. I used conservative industry averages and made it clear they’re estimates.

Another challenge was designing a reward system that encourages efficiency—not just usage volume. I focused the XP system on rewarding lower emissions per token, and highlighting greener choices.

Getting real-time data sync between the frontend and Supabase—especially for live dashboard updates and XP changes—took careful handling of state and queries.

What I’m proud of
I’m proud that Promptivore turns a fairly abstract idea—the carbon cost of AI—into something visible and manageable. The dashboard, real-world equivalents, and XP system make it easy to stay aware without needing a science degree. The backend is solid, and the app is fun to use without being patronising or pushy.

Hero Dashboard

Promptivore Dashboard – Track AI usage and emissions

What I learned
I learned more than I expected about how model choice, prompt length and usage habits affect carbon emissions. I also learned how to build a gamified system that nudges behaviour without becoming a leaderboard game. This project deepened my experience with Supabase, particularly around auth and secure session storage, and helped me level up my frontend flow with Tailwind and React.

What’s next for Promptivore – Mindful AI usage, visualised
Planned improvements:

  • Dark mode for late-night use
  • More detailed analytics and future emissions projections
  • Community features like shared badges and challenges
  • Direct API integrations to automate logging and boost accuracy
  • Expanded offset options with more detail on providers
  • Deeper gamification with unlocks and challenges
  • Improved model data based on location and hardware specifics

Additional context
Promptivore was built with Bolt.new as a showcase of what’s possible when building fast, privacy-first, full-stack web apps with modern tools.

Built With

  • bolt.new
  • chatgpt
  • netlify
  • supabase
Share this project:

Updates