4d ago · 12 min read · I spent time going deep on VoiceArena's TTS evaluation data — 44,387 human pairwise votes across six languages. What I found was not which model ranked first. It was how dramatically models collapse o
Join discussion
May 25 · 5 min read · Enterprise calls fail quietly. A caller interrupts and the system resets. A question falls outside the expected script and the bot routes to hold. These are not edge cases in most contact centres. T
Join discussion
May 11 · 16 min read · A practical May 2026 breakdown of STT, TTS, voice agent platforms, latency budgets, and cost trade-offs for production LLM voice systems. Voice agents finally feel like an engineering problem, not a r
Join discussion
May 10 · 9 min read · On May 8, 2026, OpenAI shipped three new voice models into its API — and the most significant of them changes what voice agents can actually do. GPT-Realtime-2 is the first voice model in the Realtime API family to carry GPT-5-class reasoning. That c...
Join discussionMay 10 · 9 min read · On May 8, 2026, OpenAI shipped three new voice models into its API — and the most significant of them changes what voice agents can actually do. GPT-Realtime-2 is the first voice model in the Realtime API family to carry GPT-5-class reasoning. That c...
Join discussionMay 10 · 14 min read · Key Takeaways OpenAI shipped three new audio models on May 7, 2026: GPT-Realtime-2 (GPT-5-class reasoning for live voice), GPT-Realtime-Translate (70 input languages, 13 output languages), and GPT-Realtime-Whisper (streaming speech-to-text). Pricing...
Join discussion
May 10 · 9 min read · On May 8, 2026, OpenAI shipped three new voice models into its API — and the most significant of them changes what voice agents can actually do. GPT-Realtime-2 is the first voice model in the Realtime API family to carry GPT-5-class reasoning. That c...
Join discussionMay 5 · 16 min read · OpenAI's May 4 engineering post on voice AI at scale. Notice what's missing from the headline. Key Takeaways OpenAI published a 4,000-word engineering post on low-latency voice AI on May 4, 2026, and didn't include a single millisecond figure in the...
Join discussion
May 2 · 9 min read · The text-to-speech market has been locked behind proprietary APIs for years. ElevenLabs charges per character. OpenAI TTS requires an API key. Every product you build hands your audio pipeline — and your users' data — to a third-party server you do n...
Join discussion
