Deepgram announced on Thursday that it is acquiring Daily, the company that maintains Pipecat, an open-source voice AI infrastructure. The companies did not disclose the deal amount but confirmed that Pipecat will remain open-source. Pipecat functions as a framework managing voice and video streams alongside components such as speech-to-text (STT), large language models, and text-to-speech (TTS) models while minimizing latency.
With this acquisition, Deepgram is expected to gain a deeper understanding of real-time voice and video infrastructure. Daily has secured over $62 million in funding from investors including Tiger Global, Highfield Capital, Nat Friedman, Lachy Groom, and Kevin Weil, according to PitchBook data. Deepgram itself reached unicorn status earlier this year following a $130 million Series C round.
Kwindla Kramer, Daily’s co-founder, stated that the company began experimenting with AI-to-human communication over voice and text around the time GPT-4 was released, which led to the Pipecat framework. Kramer noted that he favored Deepgram’s CEO Scott Stephenson’s commitment to keeping Pipecat open-source forever, citing the matching engineering culture and the commercial advantage of keeping the ecosystem as open as possible. Scott Stephenson stated that Deepgram has long sought involvement in open source and identified an orchestration framework for agents as a potential area, driving the decision to partner with Daily.
Stephenson added that the acquisition allows developers to build agents with expressivity and end-to-end speech systems by connecting different models and a context bus. Furthermore, Kramer stated that the team viewed Pipecat as a toolkit for building new kinds of AI-native software, rather than solely a voice toolkit. Deepgram is also launching the 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 the audio, instructions for a language model on how to use those signals, and guidance on the desired sound of the response.
