Dubverse

Voices of Tomorrow: The Dubverse Podcast Experience! Which podcast do you binge over, tell us in the comments

✨ Wishing you a Diwali full of light, laughter, and limitless possibilities! ✨ This year, we’re bringing Diwali greetings straight to you through the magic of AI—our little spark of innovation! Just like the diyas that light up the darkest corners, AI is here to brighten your journey with creativity, speed, and next-gen tech. Let’s celebrate a year of development as swift as a Diwali rocket! 🚀 Happy Diwali from all of us at Dubverse🌟 #HappyDiwali #AIInnovation #FestiveTech

The Creator Economy 2025 Welcome to a bold new chapter for creators! Creators are launching product lines, becoming software founders, and serving as the new TV channels people can’t wait to tune into. They’re going global instantly and becoming even more niche. They’re getting smarter using AI and looking for more protection from those tools. Our fourth annual Creator Economy Report brings you the latest and greatest about the sector that’s overturning traditional media as we know it. Though fewer startups are now building in the creator space, the creator market size is expected to double by 2027 with intense focus on AI, community, and diversifying revenue streams. Read the latest report to discover 👇 🤔 What is the creator economy? ⏰ The transformation of the creator economy over the past twelve months 🚀 The latest 180+ startups to watch 💰 The fundraising and unicorns landscape ❓ The impact of AI 🧲 Challenges for creators today 🔍 Trends for 2025 🧙 How creators can thrive in 2025 👀 Is the creator economy here to stay? And more! This year’s edition is brought to you by: Linktree, Mighty Networks, Everbloom and KLIQ. A big thank you for powering this report. Discover how the creator economy develops from here 👉 https://lnkd.in/eucF3WcG

Dubverse Drop: Releasing Ultra Low Latency APIs for TTS! 🚀 After talking and deploying to 100s of startups and builders in the voiceagent space, we realized there's an immense need for low latency TTS solutions. We've already deployed AI voices solutions for 2M+ users across enterprise, creators, and more. Now we bring the low latency module to our AI voices, which is set to change the game for voicebots, voice agents, and beyond! Current solutions aren't serving the market effectively, and a few global players think they have a monopoly. With a $1T disruption happening in the call center industry, why should US companies dictate the price point of our AI solutions? 🔬 What we've launched: - Ultra-low latency: ~200ms response time, perfect for natural conversations - Scalability: Battle-tested in high-volume production environments - Flexibility: Streaming-enabled for real-time applications - Linguistic diversity: Robust support for multiple languages, with a special focus on Indian languages 🇮🇳 💻 Calling all innovators: Our dev-friendly APIs are designed for seamless integration. Whether you're crafting IVR systems, virtual assistants, or the next big thing in voice tech, we've got the tools you need. As AI adoption increases across the board, speech will become the default interface for interacting with any digital system. That's what we at Dubverse are building towards. 🤝 Build with us: As speech experts, we're excited to collaborate with developers, startups, IVR companies, brands, and service providers. Happy to chat and support voice based product and solutions. cc: Anuja Dhawan Jaskaran Singh Harsh Kapoor Raghav Prabhakar Abhinay Poloju Aman Chaudhary Harshil Mehta Amartya Roy Chowdhury Abhishank Saran Karan Agastya API Docs down below 👇

The Ultimate Audiobook Experience with Dubverse 🎧✨ Just a story away from escaping the fast-paced world. Immerse your audience with our new AI voices, powered by Dubverse technology, and elevate your game to the next level. 🌟📚 Come try 40+ new emotive voices launched to fuel your journey!

The future of News: AI takes the centre Stage! Explore our platform and see how our cutting-edge voices can elevate your projects — from voice bots to audio series, entertainment, and beyond. Your next breakthrough in communication is just a few clicks away. 🚀 Try our API today and bring your ideas to life with the power of AI.

🚀 𝗘𝘅𝗰𝗶𝘁𝗲𝗱 𝘁𝗼 𝗮𝗻𝗻𝗼𝘂𝗻𝗰𝗲 𝘁𝗵𝗲 𝗹𝗮𝘂𝗻𝗰𝗵 𝗼𝗳 𝗗𝘂𝗯𝘃𝗲𝗿𝘀𝗲'𝘀 𝗻𝗲𝘄 𝗧𝗲𝘅𝘁-𝘁𝗼-𝗦𝗽𝗲𝗲𝗰𝗵 𝗺𝗼𝗱𝗲𝗹𝘀! Introducing 𝗡𝗲𝗼-𝗼𝗻𝗲 ⚡ 𝗮𝗻𝗱 𝗖𝗮𝗻𝗱𝘆-𝗧𝘄𝗼 ⭐, designed to serve different needs. If you're looking for voicebot 🤖 use cases where latency matters more than quality, 𝗡𝗲𝗼-𝗼𝗻𝗲 is the solution. But if you're after high-quality output 🙆♂️, even at the expense of latency, 𝗖𝗮𝗻𝗱𝘆-𝗧𝘄𝗼 is your best bet. 𝗖𝗮𝗻𝗱𝘆-𝗧𝘄𝗼 is a next-generation TTS system trained on 𝟱𝟬𝗸 𝗵𝗼𝘂𝗿𝘀 of multilingual data. It’s inspired by 𝗧𝗼𝗿𝘁𝗼𝗶𝘀𝗲-𝗧𝗧𝗦, and earlier this year we released an open-source version of it, called 𝗠𝗮𝗵𝗮𝗧𝗧𝗦. Now, with 𝗖𝗮𝗻𝗱𝘆-𝗧𝘄𝗼, we’ve made significant improvements, including switching to more efficient 𝗚𝗲𝗺𝗺𝗮 and 𝗙𝗹𝗼𝘄 models, and leveraging 𝘄𝗮𝘃𝟮𝘃𝗲𝗰𝟮 centroids. 🔬 A few 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗮𝗻𝗱 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 𝗔𝗹𝗽𝗵𝗮𝘀: 1. Adding a 𝗻𝗲𝘄 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲 requires at least 100 hours of data. 2. 𝘄𝗮𝘃𝟮𝘃𝗲𝗰𝟮 tokens carry more than just linguistic information; they capture tonality and emotions in speech. 3. 𝗗𝗮𝘁𝗮 𝗱𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆 is key (50+ speakers) with a Gaussian distribution of clip lengths. For imbalanced datasets, load a suggested distribution during training. Data processing is an extensive task—it requires a solid pipeline using speaker diarization, voice activity detection, noise classification, and ensembled ASR systems to turn raw audio into clean training data. Once the recipe is ready, the GPUs take over. We used an 𝟴𝘅 𝗔𝟭𝟬𝟬 system to train the model, clocking in 𝟴𝟬𝗸 𝗚𝗣𝗨 𝗵𝗼𝘂𝗿𝘀 of training. Dive deeper into our research and engineering process in the full blog 🔗 in comments Big thanks to the incredible Dubverse team for all their hard work in creating a Scaled Speech Generation Model!

Excited to reveal what we’ve been building at Dubverse! 🎉 Our new voices are so real, you won’t believe it’s AI. With this latest speaker addition, we’re making a big leap across languages, bringing even more realism, scalability, and security through our APIs. What sets us apart? We’ve built our foundational models in-house, giving us full control over security and allowing us to tailor solutions that meet the specific needs of businesses. This makes our technology more adaptable and ready for the real world. As enterprise adoption of AI accelerates, it’s the perfect time to level up the same with high quality and emotion to our offerings, serving our customers' evolving needs. From voice bots to audio series and even entertainment, we’re partnering with innovative use cases to bring this AI advancement directly to the end users. With the expertise of our team at Dubverse in AI speech and voice design, we’re introducing a new era of communication. Go try the API on our platform and level up your game! 🚀 #AI #VoiceTech #Innovation #Dubverse #EnterpriseAI #MultilingualAI

About a month ago, I spent a few days at the Analytics Vidhya DataHack Summit, where I had the honor of being named one of the "Top 5 GenAI Leaders." While the recognition is appreciated, what truly inspired me was witnessing the continued evolution of tech education in India, led by the two folks standing besides me. It's remarkable to see how NIIT, under Rajendra Pawar's visionary leadership, and Analytics Vidhya, founded by Kunal Jain, have consistently been at the forefront of tech revolutions. NIIT was instrumental in equipping India's workforce with crucial IT skills during the 90s tech boom. Similarly, Analytics Vidhya has created immense value in democratizing data science education over the past decade. Now, as we stand at the cusp of the GenAI revolution, these institutions are once again rising to the challenge of shaping the next generation of tech leaders. The enthusiasm was evident - I even met a data scientist who traveled all the way from Qatar, fully sponsored by his company, to attend the summit. It's clear that there is alot of willingness to upgrade around GenAI and the larger disruption it will bring. Speaking of disruption, I'm excited to share that we at Dübverse are gearing up for something exciting. We've been heads-down developing our own foundation models in speech modality, pushing the boundaries of what's possible in speech tech and AI. Without giving too much away, let's just say we're about to make some noise in the world of multilingual speech AI. Stay tuned! Curious about the future of speech tech and multilingual AI? I'm always down for a chat about what's brewing in this space. 

We’re excited to announce a special panel discussion at the AI Founders Mixer, hosted by SenseAI Ventures and Amazon Web Services (AWS): "Building AI Startups for Global Impact" Hear from visionary founders leading the AI revolution: Panelists: Priyanka Mookerjee, Founder of CleanDirty Shivam Gupta, Founder of LLUMO AI Aseem Khanduja, Founder of Unstudio AI Varshul CW, Founder of Dübverse They’ll share their experiences, challenges, and insights on building AI startups that are making a global impact. RSVP now to join us! Registration Link: https://lu.ma/8niedqfc #aifounders #ai #SenseaAI