Introducing Universal Streaming - an ultra-fast, ultra-accurate streaming speech-to-text model for voice agents 🚀. Universal-Streaming delivers ultra-low latency, superior accuracy, and intelligent endpointing at just $0.15/hr 👇

Voice agent latency breakdown 🧵.STT: 90ms (Universal-Streaming).LLM: 200ms.TTS: 75ms.Network: 100ms+.Total: ~465ms end-to-end.🔑 Key insight: Disabling STT formatting saves precious milliseconds. Modern LLMs handle unformatted text perfectly, and every ms counts for.

🚀 Claude 4 models now available through our LeMUR API. Transform your audio into actionable insights with our industry-leading speech-to-text API --- enhanced with Anthropic's most advanced AI models. Why LeMUR + Claude 4:.🎯 Unified Solution - Speech-to-text + advanced

🩺 Build a voice-powered AI doctor that understands symptoms in real-time!. Full Next.js tutorial with AssemblyAI, Clerk auth & Neon DB. Perfect for healthcare SaaS developers 🚀.#NextJS #AIVoice #Healthcare #SaaS #WebDev .

Interesting to see all the voice agent projects people are building today 🤖.✨ AI wellness companions.📞 Customer service automation.🎓 Real-time lecture note-taking.♿ Accessibility transcription tools.🔄 CRM call integrations.The Universal-Streaming API is having its moment.

Built a production-ready voice agent in 20 minutes using Pipecat 🔥 . 🎯 AssemblyAI for enterprise-grade STT.🎯 OpenAI for reasoning.🎯 Cartesia for natural TTS.🎯 Full Docker deployment. Step-by-step tutorial 👇. #VoiceAgent #BuildInPublic.

Voice agent builders: you've optimized latency, but what about turn detection?. ❌ Silence detection: "I want. uhh. a burger" → interruption .✅ Semantic endpointing: waits for sentence completion. Universal Streaming: lightning-fast transcription + smart endpointing =.

Missed us at Zoom Developer Day? No problem!. Check out our new Zoom RTMS resource center, featuring code samples and resources for building applications that integrate Zoom with AssemblyAI 👀 .

🇪🇺 Exciting news for our European customers!. Slam-1 and LeMUR are now available through our EU API endpoint, delivering complete data residency compliance without compromising on performance. What this means:.✅ Industry-leading speech recognition with EU data residency .✅.

💡Conversation intelligence has reached a tipping point. 76% of companies now embed CI in more than half of their customer interactions. Here's what industry leaders told us:. ✅ 81% use CI for analytics & intelligence (not just sales anymore).✅ 70%+ report measurable customer.

@Vapi_AI Start using AssemblyAI today for building your robust voice agent! .

Just wrapped up an incredible voice agent hackathon with @Vapi_AI in our NYC office! 🗽🤖. 40 brilliant developers turned coffee into code and built some absolutely wild voice agents in just one day. The creativity was off the charts! Projects ranged from developer tools and

Speech-to-Text is finding its way into many applications to improve the user experience. Learn how to choose the best Speech-to-Text API for your project 👇.

RT @tom_shapland: At #CCW and interested in learning more about Voice AI agents? Join @livekit and @AssemblyAI for a happy hour at CCW tomo….

Build a voice agent with DeepSeek R1 👇.

1-day AI Voice Agent hackathon today! We can't wait to see NYC devs come build with us and @Vapi_AI! . Bring your laptop and creativity for a full day of Voice Agent fun, food, swag, and $100s in prizes. See you soon 🥳.

⏰ Final call for AI Voice Agent hackathon! . Only one more day until the @AssemblyAI + @Vapi_AI Voice Agent Hackathon in NYC! Join us tomorrow to build cutting-edge voice experiences with real-time speech AI. There are still a few spots left! 👇.

🚀 Universal-Streaming integrates with the leading voice agent orchestration platforms and tools, enabling easy adoption without requiring architectural changes:. ✅ Drop-in compatibility with @LiveKit, @trydaily via @pipecat_ai, @Vapi_AI, and other popular voice platforms. ✅

RT @kwindla: The team at @AssemblyAI released a new realtime transcription (speech-to-text) model today. It's fast and accurate. Very much….

🗓️ 1-day AI voice agent hackathon is this week! 🗓️ . Only 4 days until the @AssemblyAI & @Vapi_AI Voice Agent Hackathon in NYC! There's still space in the event to come build with our new Universal Streaming model. $500 grand prize, tech swag, food & fun this Friday! Register.

Residual Vector Quantization - a technique used in modern AI audio compression. Check it out 👇.