The voice AI space is gaining traction, and top companies are looking to strengthen their positions as more business models emerge through partnerships and acquisitions. Deepgram is the latest company to make such a move. The company announced Thursday that it is acquiring Daily, the maker and maintainer of Pipecat, one of the most popular open source voice AI infrastructures.
The companies did not disclose the amount of the deal, but said they would keep Pipecat open source.
For the uninitiated, Pipecat is a framework that manages audio and video streams, along with other parts of the pipeline such as speech-to-text (STT), large-scale language models, and text-to-speech (TTS) models, while reducing latency. Its manufacturer, Daily, provides hosted infrastructure for all types of voice and video calls.
Using Pipecat, Deepgram will gain a deeper understanding of real-time audio and video infrastructure.
Dailey has raised more than $62 million in funding to date from investors including Tiger Global, Highfield Capital, Nat Friedman, Lachy Groom and Kevin Weil, according to PitchBook data. Earlier this year, DeepGram achieved unicorn status with a $130 million Series C round. Incidentally, both companies were part of the same Y Combinator batch in 2016.
Kwindla Kramer, co-founder of Daily, said that around the time GPT-4 came out, the company started experimenting with AI and human communication via voice and text, ultimately leading to the Pipecat framework. While there is a lot of interest in the acquisition, he supports Deepgram CEO Scott Stephenson’s efforts to keep Pipecat open source forever, saying the engineering cultures of both organizations align and give Daily’s team the freedom to work on different parts of the stack within the larger organization.
“Scott was forever very committed to Pipecat being open source and vendor-neutral. I think there’s both a personal commitment and a commercial driver there. When you’re in the middle of an ecosystem that’s growing so fast, and so many parts of that ecosystem are truly the best of your company, it’s kind of a commercial advantage to make everything as open as possible,” Kramer told TechCrunch.
Stephenson said Deepgram has long wanted to work in open source, and the company decided to partner with Daily after realizing that an orchestration framework for agents was a potential area.
“We want to open up the ecosystem and allow anyone to build voice agents. It’s great that you’re using the Deepgram model, but we want you to use the model that best fits your use case. But if you realize there’s room for us to catch up in certain areas of our own models and products, that’s great. We want you to use our partners, and we’ll work to get better in those areas as well.
Stevenson said the acquisition will allow developers to build agents with expressive power and end-to-end speech systems by connecting different models and a context bus that connects all the parts.
For Deepgram, the deal also allows it to explore areas beyond voice through Pipecat, as the framework supports real-time communication in a variety of formats.
“We never thought of Pipecat as a voice toolkit. We thought of Pipecat as the toolkit we needed to build this new kind of AI-native software. Where we came from, it was always going to be real-time,” Kramer says.
Deepgram is also launching Real-Time Context Standard (RCS), an open protocol for voice agents co-developed with the Pipecat team. This protocol defines metadata such as context from audio such as whispers, instructions to the language model on how to use those signals, and guidance on how the response should sound.
There is a lot of investment in voice AI from both the model side and the infrastructure side. Model makers like Eleven Labs, Gladium, and Fish Audio have raised millions of dollars. Meanwhile, infrastructure companies like LiveKit and Vapi have raised multimillion-dollar funding rounds this year.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
