Speechify

Introducing Simba 3.5. Every language has its own way of saying "I love you." In Seoul, it often sounds like "Have you eaten?" In Moscow, it's a mother at the door: "Put your hat on, it's minus twenty out." In Tbilisi, it's a toast to the guests. In Paris, it's a baker handing over bread that's still warm. These words hide in small gestures and everyday lines, and they only touch your heart when they sound like home. That's what we built Simba 3.5 for: one text-to-speech model that speaks 48 languages like a local, out today. Every voice in this film is Simba 3.5, speaking its own language. For developers: one model id, simba-3.5, live today on every plan at half price until 21 November. What's the line that means "I love you" in your language? Watch it with sound and meet the voices: https://lnkd.in/evegdTcF

Apps built on the SpeechifyAI Text-to-Speech API can now stop one misbehaving user without slowing down anyone else. Most text-to-speech APIs limit the whole app at once: one person hammering it or misusing it, and every customer of that app pays for it in errors. Developers now send safety_identifier with each request, a hashed id for their own end user, never a name or an email. On request, we can then limit that one user or block them, while the app's other users carry on. On 13 October 2026, none of the seven other text-to-speech APIs we checked documented a way to do this. How it works: https://lnkd.in/eYZFcnvg

Two independent benchmarks in two days. Voice Arena now measures Simba 3.2 at 123 ms from sending text to hearing the first sound, the fastest on their US English board, and more than three times faster than their earlier reading. Same voices, same price, and every developer on the Speechify API got it without changing a line of code. Thank you, Voice Arena, for measuring it in the open.

Our voice AI got about 4x faster, and an independent benchmark noticed before we announced anything. Coval tests text-to-speech providers around the clock. Last week they watched our Simba 3.2 and Simba 3.0 models go from roughly 400 ms to 100 ms between sending text and hearing the first sound. Same models. The speed came from rebuilding how we serve them. Thank you, Coval, for measuring it and for the kind words.

Earlier this summer, we launched voice typing within Speechify totally for free to every user across iOS & our desktop app. No limits. Since then, we've been lucky to have enormous adoption and countless users reach out to say that they've switched over to Speechify for voice typing. The best listening experience, voice typing, and, for a growing number of users in our rollout, the ability to delegate work end-to-end to agents within Speechify. Michael Mignano recently said on 20VC that the labs will not win the app layer. We agree. Speechify is a frontier lab for voice but building intuitive, mass market AI user experiences has always been in our core DNA. AI audio & voice typing are just the beginning. – Written with Speechify Voice Typing

I'm excited to share that I've joined Speechify as General Counsel & VP of Business Strategy. Speechify is the world’s leading voice AI platform. Our latest AI voice model has been ranked No. 1 on the independent Artificial Analysis leaderboard. Speechify’s products, including text-to-speech, speech-to-text, and an API powering real-time voice agents, have reached more than 60 million users in 200 countries. I'm deeply grateful for my time at Cravath. The mentors and colleagues I worked with at the Firm have set a standard for rigor and judgment that I'll carry with me throughout my career. As the law of AI is written in real time, I could not be more excited about this opportunity to shape it from inside this incredible company. #Speechify #VoiceAI

Speechify's infrastructure was a bottleneck. Now they serve dynamic pages to 60 million users on Vercel using Next.js, Cache Components, and Instant Rollbacks. "Without Vercel, we couldn't compete at the speed this market demands." https://lnkd.in/gJWpYH8b

Vercel just published a deep dive on the website I work on every day, and it's been exciting to see some of the engineering behind it shared publicly. Over the past couple of years, I've had the opportunity to help build the Speechify website—a platform that's grown far beyond what most people think of as a marketing site. Today, it serves hundreds of thousands of dynamic pages across dozens of languages, supports an experimentation-heavy growth engine, and reaches millions of users worldwide. Some of the engineering problems we've solved include: - Scaling to 500,000+ pages across 40+ languages. - Serving highly dynamic content while keeping pages fast. - Building fully editable content and page layouts directly from Payload CMS. - Running continuous product experiments without sacrificing performance. - Leveraging Next.js caching, ISR, and Cache Components to keep infrastructure efficient at scale. One of the most rewarding parts of working on the platform has been building reusable systems instead of one-off pages. We've invested heavily in a component-driven architecture where content and layouts are fully editable through Payload CMS, enabling teams to create and iterate quickly while keeping the codebase consistent, scalable, and performant. The platform is also deeply experimentation-driven—we rely on it to continuously test ideas, validate hypotheses, and ship improvements quickly, all while maintaining a fast and reliable experience for millions of users. Huge kudos to everyone across Engineering, Design, Product, SEO, and Content. Building a platform at this scale is a true team effort, and I'm grateful to have worked alongside such an incredible group of people. If you're interested in how we built and scaled a modern Next.js platform for millions of users, the Vercel case study is well worth the read. https://lnkd.in/dZR74TbE - Written with Speechify Voice Typing