Ecologits

EcoLogits is a Python library that calculates the energy consumption and CO₂ emissions of generative AI models accessed via APIs. It supports major providers like OpenAI, Anthropic, Cohere, and more, providing a transparent view of the environmental impact of each API call. The tool is designed for developers, data scientists, and sustainability-minded teams who want to monitor and reduce the carbon footprint of their AI applications. How it works: EcoLogits intercepts API requests to supported models, estimates the energy used based on model size, token count, and hardware assumptions, then converts that into grams of CO₂ equivalent using regional grid emissions factors. It integrates seamlessly into existing Python workflows via a simple context manager or decorator, automatically logging impact data alongside API responses. What makes it different: Unlike generic carbon calculators, EcoLogits is specifically tailored for generative AI models, accounting for the unique energy profiles of transformer inference. It offers granular per-request tracking, supports real-time streaming responses, and can be used to enforce organizational carbon budgets. The project is open-source and community-driven, with a focus on accuracy and transparency. Who it's for: Developers building AI-powered applications who need to report or reduce environmental impact, sustainability officers auditing AI usage in their organizations, and researchers studying the ecological costs of large language models.