Back to nFOX

QUICK START ยท 4 min 38 sec

Stand up your first contained model.

Watch the nFOX Quick Start walkthrough to learn how to stand up a contained AI model using nFOX and the Blockhouse sandbox.

Follow the walkthrough

  1. 0:00 Introduction to nFOX Fleet Manager

  2. 0:14 Select a model from Hugging Face

  3. 0:45 Choose an inference engine

  4. 1:08 Connect your GPU environment

  5. 1:41 Enable Blockhouse sandbox containment

  6. 2:08 Scan the model surface

  7. 2:50 Hydrate the model and verify provenance

  8. 3:19 Put the model into production

  9. 3:33 Test the model with chat

  10. 3:44 Connect a chat client or internal process

  11. 3:59 Tune serving performance

  12. 4:10 Next steps and pilot support

Ready to get started?

Tell us about your AI experience and GPU setup so we can guide your private AI trial.

Start your trial Read the user guide
READ THIS FIRST!Just a few things to get the most out of your pilot!

You can send your request first. Here's what will help you get ready.

A Hugging Face token helps

Some models need permission. Others are just big. You don't strictly need a token, but it helps. A token lets nFOX access the models your Hugging Face account can use. Keep it for setup; don't paste it into this form.

How to create a token (opens in a new tab)

Arrange GPU access

We've found RunPod and Lambda easiest to set up for quick GPU access. Already have a GPU? Bring your own.

Explore RunPod (opens in a new tab)Explore Lambda (opens in a new tab)
GPU rental is billed separately by the provider.