Telnyx

Mission critical communications, voice AI at scale, and compliance that moves in days instead of months are the reasons customers give for running on Telnyx, in their own words. We asked eleven of them to finish one sentence: "You should use Telnyx if..." A CEO answered with cutting edge technology. A clinical engineer answered with control, scale, and unit economics. A VoIP founder answered with magical powers. 🪄 Eleven leaders across different industries and different products could all finish the same sentence without hesitating. That is what infrastructure you can trust sounds like. See what you can build at telnyx.com #AIInfrastructure #VoiceAI

DeepSeek V4 Flash is a production inference model. But how does it perform under real workloads? We benchmarked it across four inference providers and six prompt profiles, measuring throughput and E2E latency from 1k to 100k input tokens. The results: - 𝟭𝟱𝟲.𝟱 𝘁𝗼𝗸/𝘀 p50 throughput on Telnyx stateful, 23% ahead of Baseten - 𝟭𝟮.𝟮𝟰𝘀 p95 E2E latency, the lowest among providers with 60/60 successful runs Same model. Same benchmark. Different infrastructure. See the full results and methodology: https://dub.sh/aasq7R3

Email is the newest channel on one AI-ready communications platform. Same API key. Same webhook format. Same event stream. Now your AI agents can send a receipt via email, follow up via SMS, and escalate on a voice call, all from one backend. Learn more: https://dub.sh/JvGN97h Ask your agent to build it for you: https://dub.sh/wzu983N Watch the demo: https://dub.sh/Lb8NpEG #emailAPI #AI #developerexperience.

Telnyx CEO David Casem on the DeepFest Main Stage at LEAP David shared his take on why so many enterprise AI projects look right on paper and struggle once they meet real customers, and where the industry's default choices quietly work against the teams making them. He talked about what it costs to let the vendor you already own define your AI strategy: the familiar interface, the easy procurement, the security sign-off already in place, and the direction someone else ends up setting for you. And he made the case against the best-of-breed stack. Best model, best transcription tool, best voice platform, wired together. Sound on paper, but in production every seam adds latency and another point of failure, and your customer hears every one of them. Come continue the conversation at Hall 5, Booth H48. #LEAP26 #DeepFest #Riyadh #Saudi #KSA #MENA #AI #AIInfrastructure #AIAgents #VoiceAI #Inference #EnterpriseAI

Day 1 at LEAP isn't even over and the booth hasn't slowed down for a second! If you haven't come by yet, here's what you're missing: - Beat the Bot, our sub-500ms latency game with a prize on the line - A live voice agent handling real conversations and completing tasks in Arabic and 100+ other languages - The infrastructure it all runs on, inference, edge compute and voice AI on one platform, in region. CNBC also stopped by to interview our CEO David Casem, who took the DeepFest Main Stage this afternoon. 3 days left. Come by Hall 5, Booth H48 for a live demo. #LEAP26 #DeepFest #Riyadh #Saudi #KSA #MENA #AI #AIInfrastructure #AIAgents #VoiceAI #Inference #EnterpriseAI

Join Telnyx CEO David Casem tomorrow on the DeepFest Main Stage, Day 1 at LEAP The panel takes on "Cracking the Enterprise AI Code". What it costs when your existing vendor sets your AI strategy, and why stacks that look right on paper struggle in production. Hear David make the case for owning the infrastructure underneath your agents, and explain where the best-of-breed approach starts to break under real load. Monday 31 August, 15:00. DeepFest Main Stage. Find the team after at Hall 5, Booth H48. #LEAP26 #DeepFest #Riyadh #Saudi #KSA #MENA #AI #AIInfrastructure #AIAgents #VoiceAI #Inference #EnterpriseAI

GLM-5.3 from Z.ai. is live on Telnyx. A 753B-parameter reasoning model with 1M token context joins our roster of leading open-source models hosted on Telnyx infrastructure. GLM-5.3 has comparable intelligence to Kimi-K3, but it beats K3 on price. It's roughly half the cost of Kimi K3 on input and 67% cheaper on output, perfect for devs looking to scale agentic use without the pricetag. GLM-5.3 is running on telnyx-owned GPUs, on the same private backbone as your voice, messaging, and compute traffic. Test it today in the portal or via API https://lnkd.in/gFZZzw_k

🚀 Email just joined the conversation. 🚀 One API, one bill, one platform for voice, SMS, WhatsApp, and now email. If you are already building on Telnyx, email is an add-on. If you are not, this is a reason to consolidate. Check out the release note to learn more: https://lnkd.in/g-3H6C6a #emailAPI #omnichannel #communicationsAPI

Correct. Close. Wrong. We built a language flashcard demo around those 3 outcomes. It is a useful test for voice AI because the product has to do more than continue a conversation. It has to turn a spoken answer into feedback that helps the learner decide what to do next. Our new article explains why we chose this format, how the loop works, and the Telnyx TTS, STT, and AI Inference services behind it. We also have embedded a working video and GitHub repository of the mini-app in the article.