Friday Private Inference

Private Inference

Private inference on infrastructure you control. Connected nodes, private search, and SDK access stay on the cluster — not a cloud round-trip.

Friday

Private Inference · Nodes

Path

Node to node · private

Serving

Gemma-2-9B on-site

Links

5 nodes connected

Egress

No public cloud

Nodes

Inference nodes stay connected

Private inference runs across linked nodes on your cluster. Requests stay on the path you control — site to site, not a cloud round-trip.

  • Nodes link through Veronica Core

  • Site, edge, and air-gap together

  • Models served on-site

  • No public cloud hop

Private Search

Private Search on your own cluster

The same chat operators use in friday-rag — greeting, prompt cards, and a composer that never leaves the perimeter.

  • Four AI starter prompts

  • Agent: Normal Q&A

  • Model served on-site

  • Ask Friday anything...

Agent: Normal Q&AModel: Gemma-2-9B
GA

Hello, User

How can I help you today?

What models run on this cluster?PromptSummarize the infrastructure drawingPromptCheck BCA escape clauses on Level-1PromptWhich sources stay in Tender Alpha?Prompt
Ask Friday anything...

LLM can make mistakes. Check important info.

SDK

Connect third-party applications

Friday AI reaches customer apps, enterprise platforms, and internal tools through the SDK. Private inference stays on the cluster — this is the connectivity path, not a cloud round-trip.

  • Drop-in SDK

  • Inference stays private

  • Third-party connectivity

  • Same cluster, more surfaces

Friday