Gladia

*Life update* I’ve officially moved to New York City to lead Gladia Inc. — and I can’t wait to celebrate this new chapter with you all next week! 🇺🇸🎉 What a ride it’s been. My wife and I packed up our three kids (losing their favorite toys in the process 😅), crossed the ocean, and set off on the adventure of a lifetime — building Gladia in the US. She still thinks I’m a little crazy… but supports the madness anyway 😄 With more than 50% of our revenue (and we've just hit 300k users the other day!) coming from North America already, with its exploding voice AI market, this felt like the right move at the right time. The energy here is incredible — the curiosity, the optimism, the “let’s make it happen” attitude — it’s contagious. The mission stays the same: to turn voice into the ultimate interface. But now, we’re doing it from the heart of innovation. Next Wednesday, Constant (in the pic with me!), Anna, Thibaut, and I will host a celebration with Illuminate Financial to mark this milestone with our friends, partners, and fellow builders from the voice and AI community. Join us! 🍸 Spots are limited (sorry in advance if we run out!), but this will be the first of many events to come here in NYC: https://luma.com/sf012qna

Voice AI is entering its Mature Era, moving from experiments to enterprise-grade systems. In this conversation between Alex Casella (CTO @ Thoughtly) and our CEO, Jean-Louis Quéguiner, we covered what it really takes to build and scale production-ready voice agents: ⚙️ latency and orchestration 🧠 evals and benchmarking 🛡️ compliance and infra strategy 🚀 and how to find product-market fit in a fast-moving market In the bite-sized preview below, Alex and JL also touch on one of the biggest strategic questions for any voice AI builder: Should you self-host your models, or rely on specialized vendors? The answer is not always clear-cut, but for Alex, when it comes to voice agent builders, one thing is certain: we’ve reached the point where the market moves faster than any one startup can keep up by hosting its own models. Read the full recap or watch the replay → https://lnkd.in/eY8n33xy

From "she sells seashells by the seashore" to "the crooked wookie hooked on cookies", Gladia's Tongue Twister challenge was a hit at Vapi Con 🔥 All day we challenged builders to read the fastest, trickiest lines we could throw at them, from classics to Star Wars mad libs, while Gladia's Real-time STT tried to keep up. The result? Epic attempts, surprise victories, and a lot of laughter. Huge kudos to everyone who stepped up to the mic: the speed, accents, and improv were *chef’s kiss*. This was the perfect stress test for latency, accuracy, and multilingual transcription… and believe it, stressed we were 😅 But we're happy that together, our volunteers' and Gladia's speech chops delivered. gg Amit Navindgi Coline Labadie de Faÿ Kalpesh Shethia, ETA Certified Payments Expert 🎯 Want to see how fast-and-precise your diction is? Try our geeky tongue twisters in the Playground and put Gladia’s STT to the test. 👉 Link and list of funky phrases to try in the first comment.

Voice agents are becoming a crucial staple in making customer support dynamic, efficient, and scalable. But they also bring risk. What if they hallucinate, offer strange discounts, or break compliance rules? These are real concerns for companies. And voice agent providers must take them seriously. Our latest article details how to build voice AI systems you can actually trust, with practical guardrails including: • Retrieval-first architectures • Runtime constraints • Hallucination audits • Real-time speech filters • Voice UX that reinforces safety Read the full post 👉https://lnkd.in/gCvqzPHp

New case study just dropped 📚 This month, we take a look at how Aircall, the leading voice platform for customer communications and intelligence, cut their transcription time by 95% using Gladia. Key outcomes: • Async transcription time dropped from up to 30 minutes to under 1.5 minutes per call, end to end. • Gladia’s multilingual STT delivered production-grade reliability across Spanish, German, Italian, and more. • Aircall now runs over 1 million transcriptions per week, powering search, summaries, sentiment analysis, and agent coaching. • A clean API, strong accuracy, and low latency combined to unlock faster downstream AI features for users. 👉 Read the case study: https://lnkd.in/eRRPKAKz

Level up your voice agents in  <10 minutes 📈 We’re excited to share a new Vapi tutorial that uses Gladia’s Real-time STT: no fuss, just fast, accurate transcripts that make your AI voice assistant actually helpful and conversational - in 4 languages! Highly recommend Henryk Brzozowski's video for anyone looking to build and operationalize a fast and accurate multilingual voice agent from end-to-end (and no, we're not only saying this because Henryk refers to Gladia's STT as his workflow's 'magic ingredient' 🪄) ▶️ Watch here: https://lnkd.in/eZYJAtcG

What does it really take to build a voice agent that doesn’t just talk, but closes? Our CEO Jean-Louis Quéguiner is teaming up with Alex Casella, CTO at Thoughtly to unpack firsthand learnings from: • deploying and scaling production-grade systems, • finding the right architectures that make real-time performance possible, and • direct conversations with customers who, thanks to Thoughtly, are achieving 15x ROI. Whether you’re a developer, founder, and/or a business leader building with voice, you're not gonna want to miss this live session. 📆 Thursday, Oct 16 @ 3 pm ET / 9 pm CET Grab your spot → https://lnkd.in/exYJNckc