Skip to main content

TURN Service

Metered’s TURN Servers as a Service provide a globally distributed, enterprise-grade relay for WebRTC traffic that “just works” through NATs and strict firewalls. Traffic is automatically routed to the nearest edge, or you can pin all media to a specific region for data-residency needs. The network spans 31+ regions with 100+ edge PoPs, delivering <30 ms latency and 99.999% uptime. It’s standards-compliant (core STUN/TURN/ICE RFCs) and supports UDP, TCP, TLS, and DTLS on common ports 80/443 to maximize reachability in corporate environments.

For the detailed pricing your can visit the pricing page.

TURN URLs and Ports​

Use our global endpoints for automatic geo-routing to the nearest edge, allowing you to connect even from the strictest networks. In most cases, you should point clients to global.relay.metered.ca on ports 80 and 443 (UDP/TCP/TLS); these ports are widely permitted by enterprise firewalls.

For data-residency/compliance, you can pin traffic to a specific region using our region-specific TURN hostnames listed below.

Ports & transports (quick reference global)​

PortProtocolScheme/TransportWorks With
80UDPturn:global.relay.metered.caTURN
80TCPturn:global.relay.metered.ca:?transport=tcpTURN over TCP
443UDPturn:global.relay.metered.caTURN
443TLSturns:global.relay.metered.ca:?transport=tcpTURN over TLS

Tip: keep multiple entries (UDP first, then TCP/TLS) so ICE can fall back gracefully.

Region-specific TURN endpoints​

For data-residency/compliance or to keep media local, you can pin TURN traffic to a specific region instead of using the global endpoint. See the full list of region-specific hostnames and ports here: https://www.metered.ca/docs/turnserver-guides/turnserver-regions

Specialized High-Speed Backbone​

One unique advantage of our TURN infrastructure is a specialized high-speed backbone that links our global TURN servers. For example, if one user is in London and another in New York, each connects to their nearest TURN server. These TURN servers are then interconnected via our private, high-speed backbone—ensuring ultra-low-latency, high-throughput communication between continents.

This design dramatically reduces the time media spends in transit compared to standard internet routing, enabling:

Consistently lower latency for long-distance calls.

Stable performance even when public internet routes are congested.

High reliability for mission-critical real-time applications.

Whether you're running an international video conference, a collaborative AR/VR session, or a cross-continent multiplayer game, the specialized backbone ensures your users experience the smoothest, most responsive connection possible.

High Speed TURN Backbone

  • Global network of 31 + regions & 100 + edge PoPs with < 30 ms round‑trip latency worldwide.

  • 99.999 % uptime SLA for mission‑critical production workloads.

  • Automatic geo‑routing to the nearest PoP or pin traffic to a chosen region for data‑residency compliance.

  • Region-specific TURN endpoints – Pin your media traffic to a specific region for compliance or data locality. See the full list of region hostnames and ports.

  • Runs on ports 80 & 443 (UDP/TCP/TLS/DTLS) to bypass enterprise firewalls and deep‑packet inspection.

  • Up to 10× higher TCP throughput and 220 % faster UDP vs other providers—ideal for bandwidth‑hungry AI, 3D and live‑stream apps.

  • Powerful REST API to create/disable credentials, set auto‑expiry, and fetch per‑user usage metrics.

  • 24 × 7 support by WebRTC engineers – Get help from people who debug ICE for a living.

Limits and Throughput​

Metered’s TURN service does not impose any throughput limits per allocation. Each TURN allocation can utilize the full available bandwidth of our global backbone and edge infrastructure, ensuring optimal performance for demanding real-time applications.

  • No artificial bandwidth caps per TURN allocation.
  • Throughput is only limited by the client’s network
  • Suitable for high-bitrate use cases such as HD video, AR/VR, and large-scale live streaming.

You can scale your usage without worrying about hidden throttling or performance degradation.