Deepgram said it is acquiring Daily, the company behind Pipecat, an open source voice AI framework. The companies did not disclose the price, but they said Pipecat will stay open source. Daily also provides hosted infrastructure for voice and video calls.
Deepgram said Thursday it is acquiring Daily, the company behind Pipecat, one of the most popular open source infrastructures for voice AI. The companies did not disclose the price. They did say Pipecat will stay open source.
The deal is the latest move by a top voice AI player to strengthen its position as the sector draws heavy investment and business models take shape through partnerships and acquisitions. Model makers such as ElevenLabs, Gradium, and Fish Audio have raised millions of dollars. Infrastructure companies, including LiveKit and Vapi, closed multimillion-dollar funding rounds this year.
What Is Pipecat?
Pipecat is a framework that manages voice and video streams and the other pieces in a pipeline. Those pieces include speech-to-text models, large language models, and text-to-speech models. The framework is designed to reduce latency. Daily, which built Pipecat, provides hosted infrastructure for voice and video calls of any kind.
Daily has raised more than $62 million to date from investors including Tiger Global, Highfield Capital, Nat Friedman, Lachy Groom, and Kevin Weil, according to PitchBook data. Deepgram reached unicorn status earlier this year with a $130 million Series C round. The two companies were also in the same Y Combinator batch in 2016.
Kwindla Kramer, Daily’s co-founder, said the company began experimenting with AI-to-human communication over voice and text around the time GPT-4 came out. That work eventually led to Pipecat. Kramer said Daily drew plenty of acquisition interest, but he was drawn to Deepgram for two reasons. The first was the commitment of Deepgram CEO Scott Stephenson to keeping Pipecat open source forever. The second was the engineering culture of the two organizations, which he said matched. That match, Kramer said, gives Daily’s team room to work on different parts of the stack inside a larger company.
Will Pipecat Stay Open Source After the Deal?
“Scott was super committed to Pipecat being open source and vendor-neutral forever,” Kramer told TechCrunch. “I think there is both a personal commitment and commercial driver there. If you sit in the middle of an ecosystem that’s growing very fast, and have many parts of that ecosystem that are genuinely the best from your company, then everything being as open as possible is like a commercial advantage.”
Stephenson said Deepgram has wanted to work in open source for a long time. The company saw an orchestration framework for AI agents as a potential area to pursue, and that idea drove the decision to partner with Daily.
“We want to open up the ecosystem so everybody can build voice agents,” Stephenson said. “It’s great if you’re using Deepgram models, but I want you to be using the best models for your use case. And if Deepgram notices that we have some ground to catch up on in a certain area for our own models and products, that’s great. We want you to use partners, and then we’ll also work on becoming better in those areas.”
Stephenson said the acquisition will let developers build agents with expressivity and end-to-end speech systems. Developers can plug in different models, and a context bus connects all the parts.
The deal also gives Deepgram a way into areas beyond voice. Pipecat supports real-time communication in multiple formats, and Kramer said the framework was never meant to be limited to voice.
“We never thought about Pipecat as a voice toolkit,” Kramer said. “We thought about it as the toolkit we needed to build this new kind of AI-native software. Because of where we came from, we just always assumed real-time.”
What Is the Real-Time Context Standard?
Alongside the acquisition, Deepgram is launching the Real-Time Context Standard, or RCS. It is an open protocol for voice agents that the company co-developed with the Pipecat team. The protocol defines metadata drawn from the audio, such as a whisper. It also includes instructions for a language model on how to use those signals and guidance on how the response should sound.
This article was written with the assistance of AI.
News Factory APP - agentic news to boost your SEO & AEO.