PlayAI's $21M Funding and The Release of a New Multi-Turn Speech Model

We've raised $21 Million to craft the voice of intelligence; delightful easy-to-use voice AI interfaces and models for developers, creators and businesses.

We're thrilled to announce that PlayAI has secured $21 Million in seed and pre-seed funding from Kindred Ventures, Y Combinator, 500Global, Race Capital, Soma Capital, Pioneer Fund and other incredible partners. This marks a pivotal step in our journey to transform how humans and machines interact, advancing human-computer interaction with intelligent, delightful, and easy-to-use voice interfaces for individuals and businesses.

In the era of large language models (LLMs), we believe that voice is the most intuitive and human medium for communication in natural language. Our mission is to make voice-first interfaces as seamless and responsive as a conversation between two people — breaking down barriers between humans and technology.

Over the last few months we have released Play 3.0 mini, a fast, lightweight, multilingual realtime Text-to-Speech model with state-of-the-art voice cloning, a no-code voice agent platform and realtime APIs for conversational voice agents.

As part of this announcement, we're proud to introduce our latest model Play Dialog, a state-of-the-art multi-turn speech model trained on hundreds of millions of conversations. Unlike traditional speech models, Play Dialog excels in understanding context and responding with nuanced emotion, enabling an unparalleled human-like conversational experience.

You can try this model in action now at our newly released Playnote app, which is like NotebookLM but with any voice, custom prompts and API access.

Checkout Play Dialog API to start integrating it in your applications.

Human Preference Benchmarks rate Play Dialog as the best Conversational Speech Model

In blind testing conducted by the independent firm Podonos, every 7 out of 10 participants preferred Play Dialog over other state-of-the-art voice models in both single-turn (text-to-speech) and multi-turn (speech-to-speech) scenarios.

In the above experiment we took a neutral and cloned it using Play Dialog and other leading TTS models, generated emotional dialogs and blind tested with humans.

Here are a few samples of the comparison.

Join us in shaping the future

This funding will fuel our vision to craft the voice of intelligence and be the cornerstone of voice-first human-computer interaction., It will also accelerate research and development, expand our global language and dialect coverage, and enable even more businesses to integrate natural voice communication into their products.

We'd love for you to join us on this journey to redefine how humans and machines connect. We're hiring across research, product and developer success roles to build the next chapter of conversational AI. Please see our open roles here - careers at PlayAI

Stay tuned for more updates, and thank you for being part of the PlayAI story.

— PlayAI Team