Deepgram is acquiring Daily, the company behind Pipecat, bringing an open-source voice-agent framework and real-time communications infrastructure under one roof. It also announced a new way for voice AI systems to share context across models.
Daily handles live audio and video connections. Pipecat coordinates speech recognition, language models and voice synthesis, including interruptions and turn-taking. Deepgram says Pipecat will remain free, open source and vendor-neutral, with an independent Pipecat Foundation being established.
The companies also introduced the Real-Time Context Standard (RCS), an open protocol designed to preserve signals such as whispers, speaker changes and background sounds as audio passes through an agent. It also carries instructions for how models should use that context.
Deepgram plans to combine Daily's network, Pipecat and its speech models in a forthcoming Voice Agent Platform. Daily co-founder Kwindla Kramer will join Deepgram as Head of Open Source.
The upside is an easier path from prototype to production. The test is whether developers can still pick their preferred models, transports and tools without being steered toward one vendor.
Get stories like this in your inbox every week. Subscribe to The Week in Perspective.