Deepgram Buys Daily, Keeps Pipecat Open Source and Launches RCS
Deepgram is buying Daily, gaining open source Pipecat infrastructure and launching RCS to help developers build low latency, model neutral voice agents.
Summary
On Thursday, October 29, 2026, Deepgram announced it will acquire Daily, maker and maintainer of Pipecat, with the price undisclosed. CEO Scott Stephenson pledged Pipecat will remain open source and vendor neutral forever. The framework orchestrates real time voice and video, speech to text, large language models and text to speech while reducing latency; Daily provides hosted calling infrastructure. Deepgram expects deeper real time expertise, expansion beyond voice and support for expressive, end to end agents built with interchangeable models and a shared context bus.
Daily raised more than $62 million from Tiger Global, Highfield Capital, Nat Friedman, Lachy Groom and Kevin Weil. Deepgram became a unicorn after a $130 million Series C earlier in 2026, and both companies joined the same Y Combinator batch in 2016. Daily cofounder Kwindla Kramer said experiments around GPT-4 produced Pipecat; despite other acquisition interest, he chose Deepgram for its open source commitment, compatible engineering culture and scope for Daily’s team to work across the stack.
Deepgram is also launching the Real-Time Context Standard, an open protocol co-developed with Pipecat that encodes audio context such as whispers, tells language models how to use those signals and guides response delivery. Stephenson said developers should choose the best models, including partners where Deepgram must improve. Investment continues across voice AI, with ElevenLabs, Gradium and Fish Audio raising millions for models, while LiveKit and Vapi secured multimillion-dollar infrastructure rounds in 2026.
Positives
- Pipecat will remain open source and vendor neutral forever under Scott Stephenson’s stated commitment.
- The acquisition gives Deepgram Pipecat’s real time voice, video and multimodel orchestration capabilities.
- Developers can combine interchangeable models with a context bus to build expressive, end to end speech agents.
- Real-Time Context Standard can carry cues such as whispers into model instructions and response delivery.
- Daily’s team gains scope to work across more of the technology stack within Deepgram.
Risks & concerns
- The acquisition price and other financial terms were not disclosed.
- Deepgram provided no closing timetable or detailed integration plan for Daily and Pipecat.
- Scott Stephenson acknowledged Deepgram may need to catch up in some model and product areas without identifying them.
- Real-Time Context Standard launches without disclosed adoption figures or named external supporters.