Private AI

Private AI with no more guesswork.

We help you define workloads, fully size the server, and provide the exact deployment path — one Glass License powers one Glass Instance on a computer or cloud server, one installation script, complete guided deployment instructions.

😎 One Glass License. One script. One Glass Instance.

What we deploy

One server. One script. Your private AI.

Tell us which workloads, users, and agents you need. We translate those requirements into a complete server specification and deployment path — and remove the CPU, memory, storage, GPU, operating-system, and cloud-instance guesswork.

MODELS

240 documented AI model families

Terminal.Glass guides model selection for your Glass Instance. Browse 240 documented AI model families and more than 44,000 documented deployment sizes, including more than 100 Meta AI model options, on the NoCloudGPT model catalog.

GUIDANCE

Fully guided server sizing

Terminal.Glass matches model families and deployment sizes to your use case, specifies CPU, RAM, storage, GPU, and the correct AWS or DigitalOcean instance — then provides one installation script.

😎 Private AI with no more guesswork.

INCLUDED

Happy Nerds Menu and DokuWiki

Every Glass Instance includes the Happy Nerds Menu and built-in DokuWiki so teams can maintain private documentation alongside their private AI deployment.

Glass Licenses

Sunrise packages for every stage.

Every Sunrise account includes two Glass Licenses, so you can begin with two Glass Instances without purchasing a second starter account. Add Glass Licenses to a Sunrise account for $99 each.

Sunrise Starter — $199

Includes 2 Glass Licenses for small businesses and first deployments — private AI on your server or in your cloud.

Sunrise Business — $399

Includes 6 Glass Licenses for teams and multi-location businesses — expand with additional licenses for $99 each.

Agents

Glass Agents and Jet Agents.

Glass Agents powered by OpenClaw work through approved tools and workflows. Jet Agents use Ollama-powered CPU or GPU compute to run the private AI models selected for the workload.

Glass Agents powered by OpenClaw

OpenClaw-powered agents operating through customer-approved tools and workflows inside the Terminal.Glass environment.

Jet Agents — Ollama CPU/GPU compute

Ollama-powered CPU or GPU compute agents that run selected private AI models on hardware you approve within your deployment.

Model selection

240 model families. More than 44,000 deployment sizes.

Terminal.Glass helps you choose model families and sizes that fit your Glass Instance and workload. Research the full catalog on NoCloudGPT, then we size and deploy the right choices on your infrastructure.

Guided model selection

Tell us your use case. We match model families and deployment sizes to your CPU, RAM, storage, GPU, and cloud instance — so you deploy what fits, not what sounds impressive.

Model catalog on NoCloudGPT

Browse 240 documented AI model families and more than 44,000 deployment sizes on the NoCloudGPT model catalog — then bring your shortlist back for guided Glass Instance sizing.

Deployment

NoCloudGPT, AWS, and DigitalOcean.

Run Terminal.Glass on an Ubuntu Linux 24/26 virtual machine or standalone server, a customer-owned AWS instance, or DigitalOcean — with Google Cloud Platform and Microsoft Azure support coming soon.

NoCloudGPT — Ubuntu Linux VMs and standalone servers

Run Terminal.Glass on an Ubuntu Linux 24/26 virtual machine or standalone server you control — completely private when inference stays on your hardware.

AWS and DigitalOcean

Customer-owned cloud accounts with Terminal.Glass recommending the appropriate instance, matched to selected models and workloads, with one-script deployment.

Order

😎 Private AI with no more guesswork.

You create the recommended server, run one installation script, and receive one complete Glass Instance. Deployment documentation and available support options are provided with your account.

1

Choose hosting

2

Define workloads

3

We deploy

4

Your team starts