Why We Built GLACIS on Cloudflare Workers

We built GLACIS on Cloudflare Workers because infrastructure choices matter. AI assurance cannot add hundreds of milliseconds to every inference. It needs to stay close to the request path, and that pushed us toward an edge runtime.

Cloudflare Workers

Edge runtime infrastructure

The Latency Problem in AI Compliance

Traditional compliance infrastructure wasn't built for AI. Security questionnaires, annual audits, and policy documents work fine when you're evaluating software that ships quarterly. But AI systems make thousands of decisions per second, each one potentially subject to regulatory scrutiny.

If your attestation layer adds 200ms to every API call, you've just made your AI unusable in production. Healthcare providers won't wait. Financial traders definitely won't. We needed infrastructure that could generate cryptographic proof in single-digit milliseconds.

<50ms

Global attestation latency

330+

Cloudflare network cities

<5ms

Workers cold-start target

Why Edge Computing Changes Everything

Cloudflare says its network spans more than 330 cities and places compute, storage, and security within 50 milliseconds of 95% of internet users. We use that distribution to keep attestation close to the request path instead of routing everything through a single region.

Traditional Cloud

Centralized data centers
100-300ms latency globally
Cold starts on serverless
Regional compliance complexity

Cloudflare Edge

330+ network cities
Compute and security within <50ms of 95% of internet users
Workers cold-start target under 5ms
Regional controls can support data-sovereignty architectures

This architecture is not just about speed. When data-localization controls are configured correctly, edge processing can reduce unnecessary cross-region traffic. When attestation happens close to the request path, audit trails are easier to correlate with the underlying transaction.

Security at the Infrastructure Level

Building on Cloudflare means building behind a platform that offers managed security tooling at internet scale. Our attestation endpoints sit behind the same classes of controls that enterprises already use for public-facing applications:

DDoS Protection

Cloudflare publishes network-scale security and availability data for its edge. The practical takeaway for us is straightforward: our attestation APIs sit behind infrastructure designed for large-scale abuse and traffic spikes.

Web Application Firewall

Every request to GLACIS passes through Cloudflare's WAF, blocking SQL injection, XSS, and other OWASP Top 10 vulnerabilities before they reach our code. Healthcare and financial services clients expect this level of protection.

Zero Trust Security

Cloudflare Access provides identity-aware access controls. Combined with our cryptographic attestation, we can prove not just what happened, but who authorized it—essential for audit trails in regulated environments.

What This Means for AI Teams

When you integrate GLACIS for continuous attestation, you're not adding a slow compliance layer to your AI stack. You're adding:

Invisible latency: Attestation that doesn't slow down inference
Global resilience: anycast routing and managed failover on Cloudflare's edge
Regional control options: architectures that can be configured around data-sovereignty requirements
Enterprise security: DDoS protection, WAF, and bot management included
Cryptographic proof: Hardware-backed attestation at edge locations

For healthcare AI vendors: the HIPAA Security Rule includes audit controls at 45 C.F.R. § 164.312(b). With GLACIS on Cloudflare, the goal is tamper-evident logs generated close to the transaction path without degrading the user experience.

Why the Platform Choice Matters

The main customer-facing benefit of our Cloudflare architecture is not the startup-program story. It is that the underlying platform is designed for globally distributed traffic, edge execution, and managed security controls.

For our customers, this translates to:

Confidence that our infrastructure scales with your AI deployment
Access to a mature edge platform with continued public investment
An architecture built around distributed execution rather than centralized compliance hops

Primary Sources

Building the Future of AI Compliance

Regulatory frameworks for AI are evolving rapidly. The EU AI Act is now in force. State-level laws in Colorado and elsewhere are creating new requirements. Healthcare, financial services, and enterprise buyers are demanding proof—not just promises.

Traditional compliance infrastructure can't keep up. Annual audits don't work when AI systems make millions of decisions between reviews. Documentation doesn't prove that guardrails actually executed.

We built GLACIS on Cloudflare because we believe AI compliance needs to be continuous, cryptographic, and fast enough to be invisible. The edge is where that becomes possible.

Why We Built GLACIS on Cloudflare: Global Latency, Edge Compute & Security