# AINode: Local AI platform for NVIDIA GPUs. Chat, API, fine-tuning, multi-node. One command.

> AINode turns an NVIDIA GPU box into a complete local AI platform. Browser chat with measured throughput, an OpenAI-compatible API where one master fronts the whole fleet, a model catalog with proven launch recipes, an in-browser benchmark, quantization and single-node fine-tuning, and multi-node discovery with cross-node tensor parallel over NCCL and RoCE. Apache 2.0, free, one command per box.

## Pagine principali

- [Models tested on AINode · AINode](https://ainode.dev/models)
- [AINode model catalog · AINode](https://ainode.dev/catalog)
- [AINode Pocket · AINode](https://ainode.dev/pocket)
- [Coding harness results · AINode](https://ainode.dev/harness)
- [https://ainode.dev/.agent](https://ainode.dev/.agent)

## Sitemap

- [Sitemap XML](https://ainode.dev/sitemap.xml)
