Deepgram Agrees to Acquire Daily to Integrate Pipecat Framework into Voice AI Platform
Voice artificial intelligence infrastructure provider Deepgram announced plans on October 29, 2026, to acquire Daily, creator of the open-source Pipecat agent orchestration framework, according to reporting by TechCrunch and corporate announcements at the Deepgram Speak conference in San Francisco. Financial terms of the transaction were not disclosed by the participants. Leadership from both firms stated that Pipecat will continue under its current BSD V2 license as free and open-source software, and the companies announced plans for an independent Pipecat Foundation.
Venture Trajectory and Executive Integration
Daily previously secured more than $62 million in venture funding from institutional backers including Tiger Global, Highfield Capital, and prominent angel investors Nat Friedman, Lachy Groom, and Kevin Weil. Deepgram achieved unicorn valuation earlier this year after completing a $130 million Series C financing round. Both startups participated in the Y Combinator accelerator in 2016. Following the completion of the transaction, Daily's personnel will integrate into Deepgram's organization, with Daily co-founder Kwindla Hultman Kramer assuming the role of Head of Open Source. Kramer explained that following the arrival of GPT-4, Daily began exploring speech interfaces that eventually coalesced into Pipecat. While several prospective acquirers approached the startup, Kramer favored Stephenson's pledge to maintain Pipecat as an open and vendor-neutral resource forever. Stephenson noted that Deepgram had pursued an entry into open-source software for an extended period, identifying an agent orchestration layer as the natural fit.
The Pipecat repository currently shows approximately 16.9k stars and 3.0k forks.
Orchestration Architecture and Network Infrastructure
The acquisition combines Deepgram's automated speech recognition and voice synthesis systems with Daily's real-time communication network. Daily maintains distributed global communications infrastructure encompassing 75 points of presence engineered for low-latency live media delivery. The Pipecat framework provides client coordination that connects speech-to-text models, large language models, and text-to-speech pipelines. Deepgram Chief Executive Scott Stephenson emphasized that maintaining vendor neutrality remains essential for developer adoption, enabling engineering teams to combine disparate third-party AI models across production deployments.
Open Standards and Platform Ecosystem
At the Speak developer conference, Deepgram also introduced the Real-Time Context Standard, an open protocol developed in partnership with the Pipecat team. The standard defines structured metadata capturing acoustic nuances such as vocal inflection, whispering, and ambient background conditions, allowing language models to interpret emotional cadence and generate appropriately calibrated spoken responses. Deepgram Chief Product Officer Ed Anuff previewed the forthcoming Voice Agent Platform alongside a Global API designed to improve connectivity. Deepgram currently provides speech AI infrastructure to more than 5,000 corporate clients, serving 650,000 developers across global markets.