Deepgram Expands Its Reach with Acquisition of Pipecat

The voice AI industry is experiencing significant growth, with leading companies striving to solidify their market positions through strategic partnerships and acquisitions. Deepgram has recently announced its intention to acquire Daily, the developer and maintainer of Pipecat, a widely used open-source voice AI infrastructure. While the financial details of the transaction remain undisclosed, both companies have confirmed their commitment to keeping Pipecat open-source.

For those unfamiliar, Pipecat serves as a framework that efficiently manages voice and video streams, integrating various components such as speech-to-text (STT), large language models, and text-to-speech (TTS) models, all while minimizing latency. Daily specializes in providing hosted infrastructure for diverse voice and video calling applications. With the acquisition of Pipecat, Deepgram is expected to enhance its understanding of real-time voice and video infrastructure.

To date, Daily has successfully secured over $62 million in funding from notable investors including Tiger Global, Highfield Capital, Nat Friedman, Lachy Groom, and Kevin Weil, as reported by PitchBook data. Earlier this year, Deepgram achieved unicorn status following a $130 million Series C funding round. Interestingly, both companies were part of the same Y Combinator cohort in 2016.

Kwindla Kramer, co-founder of Daily, noted that the company began exploring AI-to-human communication via voice and text around the release of GPT-4, which ultimately led to the development of the Pipecat framework. He expressed that, despite a surge of acquisition interest, he was particularly drawn to Deepgram’s CEO Scott Stephenson’s dedication to maintaining Pipecat as an open-source project indefinitely. Kramer emphasized that the engineering cultures of both organizations aligned well, allowing Daily’s team the flexibility to innovate within a larger framework.

“Scott was deeply committed to ensuring Pipecat remains open-source and vendor-neutral forever. I believe there is both a personal commitment and a commercial incentive behind this. Being positioned in a rapidly expanding ecosystem, with many components that are genuinely the best from your company, makes openness a significant commercial advantage,” Kramer stated, according to hinews.

Stephenson remarked that Deepgram has long aspired to engage with open-source initiatives and recognized the potential of an orchestration framework for agents, which motivated the collaboration with Daily. “We aim to broaden the ecosystem so that everyone can create voice agents. While it’s beneficial if you utilize Deepgram models, I want you to select the best models for your specific needs. If Deepgram identifies areas where we need to improve our models and products, that’s a positive outcome. We encourage the use of partner solutions, and we will also strive to enhance our offerings in those domains,” he explained.

According to Stephenson, the acquisition enables developers to create agents with expressiveness and comprehensive speech systems by integrating various models and a context bus that links all components. For Deepgram, this agreement also opens avenues beyond voice technology through Pipecat, as the framework supports real-time communication across multiple formats. “We never viewed Pipecat merely as a voice toolkit. We considered it the essential toolkit required to develop this new category of AI-native software. Given our origins, we have always assumed real-time capabilities,” Kramer added.

Deepgram is also introducing the Real-Time Context Standard (RCS), an open protocol for voice agents co-developed with the Pipecat team. This protocol outlines metadata such as contextual cues from audio, including whispers, instructions for language models on how to interpret these signals, and guidance on the desired tone of responses.

The voice AI sector has attracted substantial investments from both model developers and infrastructure providers. Companies specializing in model creation, such as ElevenLabs, Gradium, and Fish Audio, have raised significant capital. Meanwhile, infrastructure firms like LiveKit and Vapi have also secured multi-million-dollar funding rounds this year.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

spot_img

More like this