Rush Overview

What is RUSH?

RUSH by Amantra is “AI agents in a box”: pre-built AI agents, a fine-tuned small language model, and a compact hardware unit that runs inside your own building. Its promise is simple: switch the token meter off. There are zero GPU costs, zero token costs, and total data security.

The problem RUSH solves

AI costs rise, security is at risk, and ROI gets questioned.

• The token meter never stops: Every prompt, task, and token adds to the cost, making AI spend harder to predict.

• You spend more, but do you get more?: As adoption grows, connecting rising usage costs to business outcomes gets harder.

• Your trade secrets leave your walls: Proprietary knowledge can enter external AI systems, putting intellectual property at risk.

The RUSH approach

AI should cost like infrastructure, not like headcount. RUSH is one fixed platform fee that absorbs more work every month. It brings together three things organisations have traditionally bought separately:

• Your agents: Choose 5–10 agents from a marketplace of 1,000+ pre-built, pre-tested, governed agents.

• Your model: A fine-tuned small language model (SLM) built for your business context, documents, rules, and language.

• One hardware unit: A compact on-premises system, running on PittsOS®, that hosts the model and agents.

How it works

1. Choose: Pick 5–10 agents that match the work costing your organisation the most.

2. Scope: Size the hardware to expected volumes and confirm which systems the agents need to work with.

3. Train: The model learns your business context within your organisation’s boundaries. Data never leaves your environment.

4. Run: Deploy the hardware and launch the agents. The cost stays flat month after month.