Slavko Babić
On-device AI · WebLLM

Local Tesla AI Assistant

This AI assistant runs entirely inside your browser. Your conversations never leave your computer.

Before using the assistant, choose an AI model to download. This download happens only once and the model is stored locally for future visits.

  • 100% privateNo backend, no accounts. Nothing is sent to any server.
  • One-time downloadThe model you choose is downloaded once and cached.
  • Reused on future visitsStored in your browser (IndexedDB) and loaded instantly next time.
  • Grounded answersBuilt from a local knowledge base of 1,907 passages sourced from Wikipedia.

1 · Choose a model

2 · Requirements & performance

Selected model

Select a model above to see its explanation and requirements.

Performance note

While the model is loading and generating answers it uses your GPU, CPU and a large amount of RAM. This can temporarily slow down the browser and other applications — most noticeably on machines with limited memory or integrated graphics. The model is released from memory when you delete it or close the tab.

Also needs: a WebGPU browser (Chrome/Edge 113+ or Safari 18+) on Windows 10/11, macOS 13+, or Linux with Vulkan.

Select a model above to continue