About NERVI

NERVI is an open-source, premium local CLI LLM client built for Intel NPU hardware. It was created to make running large language models locally as seamless and fast as possible — no cloud, no latency, just raw hardware-accelerated inference.

👨‍💻 About the Developer

Hi, I'm Durga Sai — a developer passionate about local AI inference, hardware acceleration, and building tools that push the boundaries of what consumer hardware can do.

NERVI started as a personal experiment to explore just how far Intel's Neural Processing Unit could be pushed for everyday AI tasks, and grew into a full-featured platform with extensions, a marketplace, and a clean terminal UI.

Visit my personal site: dsainvg.me

🔗 Links

• Personal Website: dsainvg.me
• GitHub: github.com/dsainvg/NERVI
• Extension Marketplace: /marketplace

🛠️ Tech Stack

• CLI: Go (with bubbletea and custom ANSI rendering)
• Backend Server: Python + Flask + Intel OpenVINO GenAI
• Model Source: Hugging Face Hub
• Hardware Target: Intel AI Boost (Neural Processing Unit)
• Docs & Marketplace: Cloudflare Workers + D1 SQLite

📜 License

NERVI is open source. Contributions, bug reports, and feedback are welcome on GitHub.