GreenPrompt

Inspiration

As large language models become part of everyday workflows, I learned about their dependence to clean water as an energy supply. Additionally, I observe how inefficient many prompts are due to repeated phrasing, unnecessary filler, and vague instructions, which increase token usage without improving output quality. I wanted to build something that helps users write better prompts while also reducing wasted compute. That led me to GreenPrompt: a tool that makes AI interactions more efficient, measurable, and engaging.

What it does

GreenPrompt is a Chrome extension with a gamified companion platform that helps users optimize prompts before sending them to AI tools. It analyzes a prompt, rewrites it into a more efficient version while preserving intent, and shows estimated savings in tokens and emissions-related impact. On the companion site, users can track their personal impact over time through badges, leaderboards, and efficiency stats, turning better prompting into a habit rather than a one-time action.

How we built it

I built GreenPrompt as a Chrome extension that detects prompt input fields and analyzes prompts before they are sent. The pipeline first checks whether the user’s request looks like a simple search query, and if so, it nudges them toward using Google Search instead of an LLM for a lightweight task. If the prompt is better suited for AI, GreenPrompt then uses regex-based task-type detection to classify the request and catch patterns like redundancy, filler wording, or missing structure. After that, Gemini performs the deeper prompt analysis and rewriting, producing a more efficient version while preserving the user’s intent, tone, and important constraints. This staged approach let me combine lightweight rule-based nudges with AI-powered optimization, making the extension both practical and computationally efficient.

Challenges we ran into

One of the biggest challenges was making sure optimization did not come at the cost of quality. Simply shortening text is easy, but preserving meaning, tone, and important constraints is much harder. I also had to think carefully about how to present impact metrics responsibly, since these are estimates rather than exact real-world emissions measurements.

Accomplishments that we're proud of

I am proud that GreenPrompt goes beyond being just a simple text shortener. It meaningfully rewrites prompts in a way that is more structured and efficient, while still keeping the user’s intent intact. I am also happy to have combined a Chrome extension with a gamified companion site, because it turns prompt optimization into more understandable statistics.

Built With

• chromeextensionmanifest
• express.js
• gemini
• html5
• node.js