Our anycast mesh spans 6 regions, 47 cities, and 190 independently operated edge nodes — purpose-built for fill-request routing at scale.
All latency figures represent 95th percentile across the past 30 days.
| Region | Nodes | Avg Latency | Coverage | Status |
|---|---|---|---|---|
| North America | 42 | 0.4ms | Operational | |
| Europe | 38 | 0.6ms | Operational | |
| Asia Pacific | 51 | 0.9ms | Operational | |
| South America | 22 | 1.1ms | Operational | |
| Africa & Middle East | 18 | 1.4ms | Operational | |
| Oceania | 19 | 0.7ms | Operational |
Every fill request traverses a purpose-built anycast routing layer optimized for sub-millisecond path selection.
Every API endpoint is advertised from all 190 nodes simultaneously. Traffic flows to the nearest node by default with automatic fallback.
Direct peering relationships with 280+ carriers and IXPs eliminate extra hops. No third-party transit for intra-region traffic.
Built-in volumetric and protocol attack mitigation at every node. Scrubbing capacity exceeds 12 Tbps across the mesh.
Health checks run every 200ms per node. Unhealthy nodes are drained within 400ms; active routes migrate transparently.
The routing engine evaluates real-time RTT, packet loss, and fill-rate scores before committing a path. Re-evaluates every 5 seconds.
All inter-node traffic is encrypted with mutual TLS 1.3. Client API connections use TLS 1.2 minimum with HSTS enforcement.
Pin your routes to specific edge nodes or regions when your architecture requires location affinity or data residency compliance.
Per-node fill-rate telemetry with 1-minute granularity. Exportable to Datadog, Grafana, or any OpenMetrics-compatible sink.
Last checked: 2025-11-12 14:23 UTC
Start with the free tier. No credit card required.