bringing ai computation to your devices while saving the planet!
ai is amazing, but its environmental impact? not so much. 😟💨
millions use ai daily, relying on energy-hungry data centers that consume vast amounts of electricity.
we asked ourselves:
✦ what if ai models could run on a decentralized network instead?
✦ what if users were rewarded for contributing computing power?
✦ what if ai usage could be eco-friendly and still powerful?
thus, greensync was born—an ai system that distributes computation across multiple devices, reducing reliance on the cloud while giving users green coins for their contributions. 💰♻️
✦ decentralized ai compute – instead of using centralized servers, greensync splits ai workloads across users’ devices.
✦ green coins – earn rewards based on your device’s contribution to ai processing. 🏅
✦ real-time impact tracking – see your carbon savings and track your contribution to a sustainable ai ecosystem. 🌍
your laptop or desktop can now be part of a global ai network—without relying on cloud servers! 🚀
✦ backend: fastapi 🏗️ handles user authentication, api endpoints, and credit calculations.
✦ distributed computing: ray ⚡ efficiently distributes ai tasks across connected devices.
✦ frontend: react + tailwindcss 🎨 provides a sleek, real-time user experience.
✦ llm processing: ollama 🧠 runs ai models locally without needing cloud gpus.
✦ data sync: firebase 🔥 keeps track of green coins, user data, and system metrics.
instead of traditional ai cloud models, we use a dynamic resource-sharing system that distributes ai tasks among connected devices. no expensive cloud needed! 💻🌱
✦ running computation in the browser was difficult due to security and performance constraints.
✦ balancing ai workloads across devices while preventing overload was tricky.
✦ keeping green coin balances accurate & real-time across the dashboard and donation pages.
✦ optimizing ray’s scheduling algorithms to maximize efficiency across different hardware configurations.
✦ ai runs without the cloud – successfully distributed ai processing without relying on cloud infrastructure. 🌍
✦ green coin system – a real-time reward system that fairly distributes credits based on contribution. 💰
✦ engaging ui – built an interactive, user-friendly dashboard that tracks earnings & impact. 📊
✦ optimized task scheduling – made ray super efficient in splitting workloads across devices. ⚡
✦ decentralized ai is challenging but possible! browser-based computing has limitations.
✦ optimizing computation sharing is crucial for performance & fairness.
✦ task scheduling with ray is key for balancing workloads efficiently.
✦ real-time ui updates require careful state management to keep everything in sync.
✦ support for more ai models – expand compatibility with bigger and better ai models.
✦ smarter task allocation – improve real-time resource distribution for efficiency.
✦ ai compute pools – allow users to team up for even larger ai projects!
✦ real-world impact – convert green coins into carbon offset donations for sustainability projects. 🌱💚
✦ backend: fastapi 🚀
✦ compute distribution: ray ⚡
✦ llm processing: ollama 🤖
✦ frontend: react + tailwindcss 🎨
✦ data sync & auth: firebase 🔥
✦ programming languages: python 🐍, javascript ⚡
💻 run the greensync server:
export ray_enable_windows_or_osx_cluster=1
ray start --head --port=6379
uvicorn server:app --host 0.0.0.0 --port=8000🖥️ connect another device as a worker:
export ray_enable_windows_or_osx_cluster=1
ray start --address=<your_ip>:6379🌐 run the frontend:
npm install
npm start🛠️ contribute to the project! fork the repo and help make ai more sustainable!
💚 support the cause! spread the word and help reduce ai’s carbon footprint! 🌍
let’s make ai greener, fairer, and more accessible together! 🌱🚀
🎨 designed with love for the planet. 🌎💚 greensync - ai that cares.