Deepgram acquires Pipecat-maker Daily to bolster its voice AI infrastructure

Deepgram, a prominent player in the voice AI sector, has acquired Daily, the creator and maintainer of the popular open-source voice AI infrastructure Pipecat. The acquisition doesn't have a disclosed monetary value, but it will stay open-source, emphasizing the commitment to openness. Pipecat is a framework that manages voice and video streams, including speech-to-text, large language models, and text-to-speech models, reducing latency in the process.

Pipecat's maker, Daily, offers hosted infrastructure for voice and video calls, and has raised over $62 million in funding led by Tiger Global, Highfield Capital, Nat Friedman, Lachy Groom, and Kevin Weil. Deepgram, which recently reached unicorn status with a $130 million Series C round, aims to bolster its voice AI infrastructure through this acquisition.

Daily's co-founder, Kwindla Kramer, explained that the company began experimenting with AI-to-human communication over voice and text following the release of GPT-4. He added that Deepgram's CEO, Scott Stephenson, was committed to keeping Pipecat open-source forever and that both companies shared a similar engineering culture, making the acquisition a great fit.

Stephenson, on the other hand, highlighted the acquisition as a way for Deepgram to work in open source and to support developers in building voice agents using the best models available.

With this acquisition, Deepgram will deepen its understanding of real-time voice and video infrastructure. Additionally, the company will explore beyond voice through Pipecat, as it caters to real-time communication in various formats. To facilitate this, Deepgram is launching Real-Time Context Standard (RCS), an open protocol for voice agents co-developed with the Pipecat team. RCS defines metadata like context from the audio, instructions for a language model, and guidance on how the response should sound.