TokenWhale
Problem Statement: Large Language Models (LLMs) are widely used in daily workflows, but every interaction with them consumes tokens, which directly translates to computational cost, energy usage, and environmental impact. As LLM usage grows, so does the demand on data centers, leading to increased carbon emissions and water consumption. Most users are unaware of how many tokens they use or how to reduce them effectively.
Idea: TokenWhale is a Chrome extension designed to help users reduce the number of tokens used when interacting with LLMs. It works by optimizing user inputs and limiting the length of model outputs, resulting in more efficient use of language models and reduced resource consumption.
Product Overview: TokenWhale runs directly in the browser and provides a lightweight interface that allows users to optimize their prompts and control the expected length of the model's response. It also tracks how many queries have been optimized to give users visibility into their usage patterns and savings.
Key Features: Prompt Optimization: Reduces input token count by rephrasing and simplifying the user’s text without altering its intent. Output Limitation: Lets users specify the desired output length (e.g., 1-2 sentences, short, medium, long, or original) to avoid unnecessarily long responses. Query Tracker: A floating window displays the number of queries optimized to keep users aware of their contribution to efficiency. Customizable Settings: Users can configure their preferences for prompt compression and output length based on their specific needs. Ease of Use: Simple to install and use with no technical background required. Integrates smoothly into the browser interface without disrupting the normal user experience. Designed to be minimal, responsive, and non-intrusive.
Log in or sign up for Devpost to join the conversation.