Hi everyone, and welcome to the first ElevenLabs Summit! My name is Mati, and me and my best friend Pot are the co-founders of ElevenLabs. Pot and I grew up in Warsaw, obsessed with technology and how fast it has advanced throughout history.
Just 20 years after the phonograph first captured sound, radio carried voices across long distances over invisible waves. Only 50 years after supercomputers required a room this size, they had moved from universities to offices to backpacks to pockets. But Pot and I recognized that while technology had developed at lightning speed, how we interacted with it had not. We type on keyboards because that's how computers need to be directed. We stare at screens because that's how they best communicate with us. We study their languages so we can code with characters they understand. And we rarely use our voices, our preferred interface, because until recently, technology had never really been taught to speak.
Pot and I also recognized the great power of technology to help us create, and the endless possibilities that teaching it to speak would unlock. It would narrow the gap between ideation and production. It would make information, content and experiences accessible to all. So on the belief that our interactions with technology should be on our terms and seamless, and that creating with technology should be seamless too, we founded ElevenLabs less than four years ago and set out to make our belief a reality. Lucky for me, Pot is the smartest person I know, so I had great faith that ElevenLabs could solve the foundational research problems and bring extraordinary products to life.
To achieve seamless interactions and creation, we had two initial problems to solve: giving technology a truly human voice, and giving it knowledge on demand. To take on the first, we formed our VoiceLab. Each lab within our company is a small but mighty collection of exceptional researchers, engineers, and operators committing to solving a problem. In this case, giving voice to technology. The VoiceLab team worked quickly. By early 2023, more than one million users were generating voices that could speak countless languages with perfect emotion, intonation, and prosody. Creators and developers, and soon after enterprises, made voiceovers at scale for ads, articles, audiobooks, and social posts. They dubbed videos and podcasts, localizing content more easily than ever before. They helped those who had lost their voice due to an accident or illness to speak exactly as they once did. All the while, the VoiceLab team never let up, optimizing inference to achieve sub-100-millisecond latency and deploying the most emotional and controllable voice models on the market today across 70 languages. Our voice models and products succeeded in empowering seamless creation.
But seamless interactions required more. No matter how good a voice is, if the infrastructure beneath it is poor, if it's disconnected from knowledge and context and intelligence, interactions won't be seamless. So we turned to that second problem — giving technology knowledge on demand — and formed our AgentLab to solve it. The team built conversational agents that power intelligent, natural, and dynamic interactions: agents that can write, listen, respond, and execute instantly. They constructed what is today our Agents Platform, with countless integrations and features that ground our agents in our customers' data, adapt them to their workflows, allow for scaling without stress, and of course incorporate our leading voice models. From Fortune 500s to the quickest-growing AI-native startups, businesses used the Agents Platform to overhaul their customer experience, forever changing how they engage with those they serve. Gaming companies created immersive experiences enabling real-time conversations with characters. Governments improved how they deliver education and public services to their people. Our Agents Platform, knowledge on demand paired with the power of the human voice, delivered seamless interactions with technology.
But for the ElevenLabs team, the mission is never accomplished. And we had identified so many other things we knew we needed to build. So we spun up new labs to solve new problems, and we started shipping faster than ever before. To name just a few: we developed Scribe, the most accurate transcription model ever, beating all the giants. We expanded into localization, elevating dubbing to new heights. We opened distribution channels with our Reader App. We released the world's best fully licensed music model, answering the call of our users who demanded an easier way to bring their ideas to life. We launched our all-in-one creative platform, bringing our foundational audio models and products together in a single place. And we've kept improving our Agents Platform, focusing on latency, reliability, and versatility, with efficient orchestration that allows for instantaneous action and the glide path from creation to testing to deployment.
Today, we serve millions, from some of the world's largest enterprises to individual creators. We've scaled globally. We've scaled our revenue from zero to almost $300 million. And we've scaled our team from two to 350 builders who care deeply about doing the best work of their lives. We are proud of what we've done, but we are just at the beginning.
So let's talk more about the future we envision and intend to make real. First, speech will be our primary interface with technology. Your laptop, wearable, smart device, car, and any machine you engage with — from charging stations to robots — will understand your intent based on what you said and exactly how you said it. With these advancements, we'll be more present, spending less time engaging with technology than keyboards and screens ever allowed. ElevenLabs will lead this transformation, meeting every customer where they are, prioritizing safety and security, delivering not just in cloud, but on-prem and on-edge. We'll also give our customers more voice control, allowing them to fine-tune and direct tone, sentiment, and cadence to fit any interaction or creation.
Second, the language barrier will fade away. Information and content will be universally accessible with real-time translation across languages and voices. And when the language barrier falls, so too do the barriers to learning, communicating, and connecting. Our localization tools will drive this change, with new models for self-serve, high-quality automated dubbing and truly perfect dubs made with humans in the loop by our productions team. Finally, the best AI products will be omni-model, omni-channel, and omni-present — designed to help solve real and specific problems with a one-stop suite of tools. Our creative platform will move beyond audio, in recognition that our customers don't just need the best of sound. They need the best of image and video, too. So we are bringing partner models to the platform to deliver a true omni experience. Our Agents Platform, already omni-channel with voice and text, will go deeper, with agents that carry knowledge and context between interactions, no matter where they occur, and with increased context and new fused speech-to-speech models, our agents will properly interpret emotions and understand you in a whole new way.
We started off with the belief that seamless interaction and creation with technology would make for a better world. Today, that belief is stronger than ever. Our small, mighty teams may tackle different problems, but we are united by our commitment to bring that belief, that one dream, to life across ElevenLabs: seamless interaction, seamless creation. And we are heads down and all in.
