Enterprise · Fly

Fly.io for Enterprises

Deploy apps near users
In 33+ regions
On 6 continents

Requests per second Worldwide
5.797e+4

Scale

Fly.io runs its own hardware. We’re not an API layer on an existing public cloud. That means we have fine-grained control over workloads, and it means we can use hardware virtualization to securely isolate apps from each other. But it also means we control our network uplinks, which is how we efficiently scale apps globally.

Support

We offer round-the-clock support with dedicated teams for customer support and infra operations. Fly.io is a distributed company operating in most of the world’s time zones. Support scales with your needs, from priority email technical support to SLAs and guaranteed response times.

Security

Fly.io is deeply committed to security. Our platform is hardened by design, with full hardware virtualization, a memory-safe stack (including a Rust hypervisor), end-to-end encryption with WireGuard, memory-safe SSH, and TLS 1.3. We have a dedicated security team with decades of software security experience at NCC Group, Matasano, Latacora, and GitHub. We do regular third-party audits. We’re SOC2 Type 2 attested and HIPAA-ready. We are not messing around with this stuff.

Infrastructure Integration

We integrate easily into CI/CD pipelines, so you can deploy automatically from git merges, including minimally-privileged deploy tokens for hosted deployment platforms like GitHub Actions. We do Prometheus metrics (and you can export your own to integrate with our feed) with dashboards and alerting, logging, and configurable health checks. All this stuff is ready out of the box, and it all integrates with the prod infrastructure you already have.

No Lock-In

The basic idea behind Fly.io is simple: we take standard OCI (Docker) containers and run them as VMs. That’s why we can speedrun most applications: if it works in a container, it works here. Our core features are built on standard protocols: HTTP/2, IPv6, WireGuard, and DNS. So if it works here, it’s going to work on the big public clouds, too. No code rewrites getting in, no code rewrites getting out.