Play

Today, we're introducing PlayHT2.0 - our new Conversational Text-to-Voice AI Model that's trained and built to generate humanlike conversations across languages with <1s latency. At the core of PlayHT2.0, is its Instant Voice Cloning capability that can capture any voice and accent from just 3s of a speaker's voice and synthesize speech in a truly conversational tone. We're also introducing, for the first time ever, the concept of Emotions to Generative Voice AI - allowing you to control and direct the generation of speech with your chosen emotion. Sign up for beta access - https://lnkd.in/gESmwwww

Founder Mode by [AI] Paul Graham https://lnkd.in/g_sNamVF ℅ Y Combinator

Our team at Andreessen Horowitz believes now is the time for AI to reinvent the phone call - enter conversational voice agents. (CC Anish Acharya) AI voice agents can navigate live calls like a human speaker - for use cases across B2B (customer services, appointments, bookings, etc.) and B2C (coaches, tutors, companions, etc.) These agents are just now possible with LLMs, and the stack is still unfolding - with players like Retell AI, Vapi (YC W21), Bland AI, and Play providing "full stack" solutions, while companies like Daily, Deepgram, and ElevenLabs allows you to build your own stack. Incoming multi-modal models like GPT-4o might also change the game here in the coming months. At the apps layer, we're excited about agents that "verticalize" for specific B2B and B2C behaviors - building out workflows, tuning models, and integrating with specific tools needed to deliver on a humanlike conversational experience. This includes companies like OpenCall.ai (YC W24), Assort Health, Happyrobot, and Hyperbound on the B2B side - and Lifelike (YC S23), Speak, and Sonia (YC W24) for B2C. Check out the market map below for what we've seen in this space so far ⬇ and our full deck for more of our voice agent thesis: https://lnkd.in/gzNz_jwB If you're building here, please reach out! 👋

Today, we're introducing PlayHT2.0 - our new Conversational Text-to-Voice AI Model that's trained and built to generate humanlike conversations across languages with <1s latency. At the core of PlayHT2.0, is its Instant Voice Cloning capability that can capture any voice and accent from just 3s of a speaker's voice and synthesize speech in a truly conversational tone. We're also introducing, for the first time ever, the concept of Emotions to Generative Voice AI - allowing you to control and direct the generation of speech with your chosen emotion. Sign up for beta access - https://lnkd.in/gESmwwww

Come join us.

Exciting update: Actions & Integrations are now available on play.ai. Seamlessly integrate custom function calls with play.ai agents and unlock new workflows powered by voice. This is the beginning in transforming agent interactions for practical utility. Here’s how it works: · Create Actions: Define your APIs, set action trigger criteria, and configure dynamic/static parameters. (Dynamic parameters are inferred by Agents at conversation time.) · Deploy to Agents: Once created, actions will be available in the Agent Editor, where you can select which actions you’d like your agent to use. Actions supercharge your agents. In the coming weeks we'll be releasing platform integrations, starting with Google Calendar. If you have a request for a particular integration, drop us a note here or in our discord community. https://lnkd.in/gWjwvt-U To see a more detailed outline on how to setup actions, check out this blog post: https://lnkd.in/gUmEKw_E Head over to play.ai to experience the magic of voice agents.

We're excited to share a new feature we’ve launched at Play.ai: Bring Your Own LLM. This feature allows developers to integrate their own OpenAI-compatible LLMs directly into our platform via our API, giving you unprecedented control and customization over the behavior and capabilities of your voice agents. Why It's Cool: Flexibility: You’re no longer limited to the built-in models; you can use any LLM that meets the OpenAI API standards. Customization: Tailor your voice agent’s responses and interactions to fit specific needs or niches. Ease of Use: Simply update the llm field through our Update Agent endpoint and you're set. We believe this opens up a new realm of possibilities for developers looking to create more nuanced and specialized voice interactions. Try It Out: You can start integrating your models today by visiting https://lnkd.in/gMEgyYyH for detailed documentation and examples. We’d love to get your feedback on this feature and discuss any ideas or questions you might have.

Play.ai by Play (YC W23) is a new real-time conversational Voice AI platform for creating human-like voice agents and interfaces. It makes speech generation contextual and handles turn-taking, interruption, voice energy, and emotion modulation for natural, fluid, and human conversations in real-time. Check out their launch at https://lnkd.in/dEHuyaMc.

Introducing https://play.ai — a new conversational Voice AI platform and API for creating human-like voice agents and voice interfaces. Powered by our new Large Dialogue Model (LDM), Play LDM 1.0 makes speech generation contextual; handling interruptions, turn-taking, voice energy and emotion modulation for a natural, fluid, human conversations in real-time. We believe natural voice is the ultimate interface for Human Computer Interaction. Human-like voice interfaces will unlock new possibilities for making AI accessible to everyone. With play.ai, you can create end-to-end conversational voice agents and deploy them to web or phone without any coding. We handle everything for you, we also provide an API if you want further customizations or integrations. Play’s API can be integrated with any LLM, AI Agent, application or device to instantly give it natural interactive conversational voice capabilities. Try https://play.ai now for free and share your feedback with us. We are working hard to make it better every day.

PlayHT (YC W23) just launched Turbo, a new blazing-fast Conversational AI Text-to-Speech model with <300ms latency. It supports input text streaming from LLMs, output audio streaming, and is able to clone any voice and accent. Try it out at https://play.ht/playground.