Run AISLE Wherever Your Code Lives: SaaS to Air-Gapped

We built AISLE with the belief that AI-native security tooling should adapt to your systems, not the other way around. After all, the organizations with the most critical security needs often have the most stringent sovereignty requirements. If you’re defending your nation, protecting your customers’ investments, or providing energy to millions of people, you shouldn’t have to give up control over your source code to access advanced reasoning capabilities. Your sovereignty and your security go hand in hand.

That’s why AISLE runs wherever your code lives. For innovators who want to move fast, we operate on a traditional SaaS model with self-serve onboarding. For those who want more control, we serve private cloud deployments. And for those who run sensitive workloads in on-prem or air-gapped environments, we send out forward-deployed engineers (FDEs) to set up AISLE with no pathways in or out.

Here’s a short breakdown of our deployment models:

	Managed SaaS	Private Cloud	On-Prem	Air-Gapped
Where it runs	AISLE-managed, public cloud (your choice of region)	Your existing cloud account	Your hardware	Your hardware, fully isolated
Models	Mix of foundation models via AISLE’s AI gateway	Bring your own: frontier and/or custom LLMs	AISLE custom models	AISLE custom models
Who sends code to model providers	AISLE, under ZDR agreements	You, through providers you choose	No one	No one
Network connectivity	Internet-connected	Connected through your own gateway	Optional outbound pull for threat and dependency intel	None: no ingress or egress
What AISLE can see	Managed instance (scope TBD)	TBD	Nothing	Nothing
Tenancy	Multi-tenant, single-tenant optional	Single-tenant by definition	Single-tenant by definition	Single-tenant by definition
Setup	Self-serve: integrates with your VCS	Self-serve: integrates with your VCS	Self-serve: integrates with your VCS	FDEs either bring code in manually or integrate with git
Who operates it	AISLE	Self-service, FDE for complex setup	Self-service, FDE for complex setup	Self-service, FDE for complex setup
Updates and intel	Continuous, managed by AISLE	Via monthly registry push/pull	Via monthly registry push/pull	Delivered manually by FDEs (cadence and verification TBD)

Managed SaaS: AISLE Runs Everything

Many AISLE customers are innovators in rapidly changing fields like crypto, data infrastructure, and biotech. They need a security platform that moves as fast as they do while enforcing strict zero-data-retention (ZDR) controls. In these cases, AISLE operates on a pure SaaS model with both multi-tenant and single-tenant deployments. You get exclusive access to a dedicated instance of both AISLE and supporting infrastructure so nothing is shared with other organizations or users.

Based on our research into the cyber capabilities of current LLMs, we plug into foundation and custom models through an AI gateway and run through the public cloud. AISLE serves customers in many regions and jurisdictions, and we’ll work with you to ensure you meet your data sovereignty obligations.

You simply connect your repositories and we fully manage the instance. Onboarding is complete in days.

Private Cloud: Your Cloud, Your Models

SaaS is convenient, but it’s not always aligned with security and compliance requirements for which models can touch your code. Some teams have substantial contracts with frontier AI labs that make it more economical to run security workloads through their existing gateway. Others may have performance, provenance, or privacy concerns with certain providers. Whatever the case may be, AISLE can accommodate a wide range of choices.

AISLE isn’t a model. It’s a system. That means you can plug it into a bring-your-own-model (BYOM) setup, with your choice of frontier and custom LLMs, within your existing virtual private cloud (VPC). We then deliver deployment artifacts in pull or push mode into your local registry.

On-Prem: Your Hardware, AISLE Models

Many critical workflows can’t be entrusted to a VPC for geopolitical, regulatory, or pure security reasons. Using AISLE, you don’t need to send your source code to a frontier lab to get advanced cyber reasoning capabilities. We constantly evaluate and fine-tune open-weight models to approach (and sometimes exceed) flagship results.

When security-sensitive customers want full authority over compute, storage, and access, we send FDEs to deliver AISLE in person. If you want threat and dependency intel, you can pull it from our servers. AISLE will never see your code, findings, or other intellectual property.

Air-Gapped: No Network Path In or Out

The most demanding operators can’t afford to have a network path in or out. For them, AISLE runs fully air-gapped using our custom models. With no network data ingress or egress, we deliver everything on physical media: signed artifacts you can verify on-site before installation, with a documented chain of custody from our hands to yours.

Both the full platform and Snapshot can run on your hardware, and AISLE FDEs can deliver updates and intel to maintain high performance without asking you to give up control. For more information about the hardware requirements for air-gapped (and on-prem) deployments, see our technical breakdown.

Security and Sovereignty Go Hand in Hand

Ultimately, sovereignty boils down to control: over your infrastructure, your data, and your tooling. Whether your sovereignty requirements are designed to satisfy a regulator, defend your national interest, or align your cost and security incentives, AISLE will work with you to help you fulfill your mission.

To see which deployment model best fits your security posture, talk to the team.