Trunnion AI, in brief.
Trunnion AI develops configurable multi-agent software and vertical products. Policy, authorization, audit, human-review, hosting, and classification capabilities are design objectives whose implementation and suitability must be verified for the exact release and customer environment.
The essentials, at a glance.
A factual summary of what Trunnion is, how it is governed, where it deploys, and how it is priced.
| What it is | An LLM-agnostic platform for configurable agent workflows with human review. |
|---|---|
| Category | Multi-agent software platform and vertical SaaS portfolio |
| Governance | Design objectives include policy-based authorization, layered tool controls, human review, and reviewable execution records. Implementation and test evidence are release-specific. |
| Deployment | Cloud, on-prem, air-gap. Designed for classification-aware work up to TS/SCI with CAC/PIV identity traceability; accreditation determined per environment. |
| LLM strategy | LLM-agnostic is an architectural objective. Provider or model changes require contractual, data-handling, security, and evaluation review for the release. |
| Products | The portfolio page distinguishes public products, development programs, and custom services. Availability is confirmed per release. |
| Industries | Defense & National Security, Government Procurement, Manufacturing & ERP, Waste & Fleet Operations, Events & Marketplaces, Real Estate, IoT & Logistics, Marketing & SEO, Travel & Hospitality |
| Pricing model | Pricing is scoped, not list-priced. What you pay depends on the product, the deployment environment, and how deep the build goes. |
| Ownership | Duskbridge is the parent holding company; the operating entity is Trunnion AI, LLC. |
| Contact | contact@viceroynm.com (9 AM to 6 PM Mountain, Monday to Friday) |
Common questions.
What is Trunnion AI?
Trunnion AI develops an LLM-agnostic multi-agent platform, vertical products, and scoped enterprise builds. Policy, audit, authorization, and human-review controls are design objectives verified for each release and deployment.
What does governed mean here?
Here, governed describes design objectives: policy-based authorization, human approval for consequential actions, and reviewable execution records. It does not assert that a specific implementation passed those controls without release evidence.
Is Trunnion tied to a specific LLM?
No. Trunnion is LLM-agnostic. A model router sends each task to the right model, hosted or on-premise, and you can swap providers without rewriting your workflows.
Can Trunnion run air-gapped?
Yes. The control plane is designed to deploy to cloud, on-premise, or fully air-gapped environments, with classification-aware workflows designed for TS/SCI-grade environments and CAC/PIV identity traceability. Accreditation is determined per environment; see the security page for boundaries.
How is Trunnion priced?
Pricing is scoped by product, deployment environment, and implementation depth, offered as per-product subscriptions or fixed-scope custom builds. The control plane itself is not licensed standalone. Request a demo for a quote.
What are the products?
The portfolio includes shipped products and separately labeled development programs across defense, procurement, manufacturing, fleet operations, events, real estate, IoT, and travel, plus scoped custom builds. See the product page for current status.
Get started
Ready to build on Trunnion AI?
See the governed control plane run against your workflow, in your environment. From cloud to air-gap.