How OpenAI delivers low-latency voice AI at scale
OpenAI News / May 4, 2026
- relay+transceiver split preserves WebRTC semantics
- server-side ufrag enables first-packet routing
- small public UDP footprint suits Kubernetes autoscaling
Matched posts: 4
Rearchitecting the Workflows control plane for the agentic era
Cloudflare / Apr 15, 2026
- 50k concurrent instances
- 300 instances/sec per account
- 2M queued instances per workflow
Introducing the Adoption news channel
OpenAI News / Mar 5, 2026
- 運用への実装
- 導入の拡大
- 信頼とガバナンス
Containers - Run 15x more Containers with higher resource limits
Cloudflare Developer Platform / Feb 25, 2026
- Memory increased to 6TiB
- vCPU limit raised to 1,500
- Disk limit expanded to 30TB
Previous1 / 1Next