# CanIRun.ai

> Detect your hardware and see which open AI models you can run locally.

CanIRun.ai grades open-weight chat, coding, image and video models against your GPU, Apple Silicon or other device. Scores use estimated speed, VRAM headroom and model quality.

## Start here

- [Can I run AI locally?](https://canirun.ai/): Best-pick models for the detected machine
- [Local AI models you can run](https://canirun.ai/models): Full filterable catalog
- [Local AI models by device](https://canirun.ai/devices): RTX, Radeon, Intel Arc, Apple Silicon, phones, Raspberry Pi
- [How detection works](https://canirun.ai/why): Hardware APIs and scoring
- [Docs](https://canirun.ai/docs): Quantization, VRAM, MoE, GGUF

## Browse by task

- [Local LLMs](https://canirun.ai/models/chat)
- [Local coding models](https://canirun.ai/models/coding)
- [Local image generation](https://canirun.ai/models/image)
- [Local video generation](https://canirun.ai/models/video)
- [Small on-device LLMs](https://canirun.ai/models/lightweight)

## Browse by VRAM

- [8GB VRAM](https://canirun.ai/vram/8gb)
- [12GB VRAM](https://canirun.ai/vram/12gb)
- [16GB VRAM](https://canirun.ai/vram/16gb)
- [24GB VRAM](https://canirun.ai/vram/24gb)
- [32GB+ VRAM](https://canirun.ai/vram/32gb)