Deepgram acquires Daily, maker of Pipecat: what the open-source voice AI deal changes

Deepgram, a voice AI company, said on October 29, 2026 that it is acquiring Daily, the company that makes and maintains Pipecat, one of the most widely used open-source frameworks for building voice agents. Neither side disclosed a price. Pipecat is to stay open source under its existing BSD V2 licence, and the companies say an independent Pipecat Foundation is being set up. For developers who build on Pipecat, the question is whether a framework owned by a speech-model vendor stays neutral about which models you plug in.

Key points

• Deepgram announced on October 29, 2026, at its first Deepgram Speak event in San Francisco, that it is acquiring Daily, the maker of Pipecat; the price was not disclosed.
• Deepgram says Pipecat stays fully open source and free under its existing BSD V2 licence, and that an independent Pipecat Foundation is being established.
• Daily co-founder Kwindla Hultman Kramer joins Deepgram as Head of Open Source, and Deepgram will sell, support and develop Daily's current products.
• Deepgram also introduced the Real-Time Context Standard (RCS), an open protocol built with the Pipecat team to carry audio context such as a whisper between models.
• Daily has raised more than $62 million, according to PitchBook data cited by TechCrunch; Deepgram raised a $130 million Series C earlier this year that made it a unicorn.

What happened

In its announcement, Deepgram says it is buying Daily "to offer the first complete, production-ready, open stack for conversational voice AI." Daily makes Pipecat and also runs hosted infrastructure for real-time voice and video calls. Deepgram describes Daily's network as providing low-latency connections between end users and voice agents; The Next Web, summarising Deepgram, puts it at 75 points of presence.

Pipecat is the piece that ties a voice agent together. A typical agent chains speech-to-text, a large language model and text-to-speech, and has to move audio streams between them quickly enough that a conversation feels natural. TechCrunch describes Pipecat as a framework that manages voice and video streams plus those pipeline pieces while reducing latency. Kramer wrote that it has more than 200 integrations and almost 400 contributors, and that 18 engineers built it, as relayed by The Next Web.

The Daily team will join Deepgram. Deepgram will "sell, support and develop" Daily's current products, and it plans to make Pipecat and Daily's network the base of a new Deepgram Voice Agent Platform, which Chief Product Officer Ed Anuff was due to preview later on the same day. Deepgram also said Anuff would introduce two new models built for live conversation, expanded language support and a Global API meant to improve connection reliability and reduce latency. Those details are the company's own plans; we have not seen them tested.

Background

Both companies went through Y Combinator in 2016, according to TechCrunch, and The Next Web reports, citing Kramer, that Deepgram has powered transcription inside Daily's products since 2019. Daily's investors include Tiger Global, Highfield Capital, Nat Friedman, Lachy Groom and Kevin Weil, according to PitchBook data cited by TechCrunch. Deepgram says more than 5,000 organisations run its speech models in production, a company figure.

Pipecat grew out of Daily's own experiments. Kramer told TechCrunch that around the time GPT-4 came out, the company started experimenting with AI-to-human communication over voice and text, which led to the framework. Voice AI has since attracted money on both sides of the stack: TechCrunch lists model makers such as ElevenLabs, Gradium and Fish Audio, and infrastructure companies such as LiveKit and Vapi, as having raised large rounds.

Kramer said several companies were interested in acquiring Daily. He chose Deepgram, he told TechCrunch, because chief executive Scott Stephenson was committed to Pipecat staying open source and vendor neutral. In the announcement he said keeping Pipecat open "was fundamental to our decision to join Deepgram."

How the open-source promise is structured

The commitments, as stated by the companies, are these:

1. Licence. Pipecat remains under its existing BSD V2 licence, free to all developers.
2. Foundation. The companies say they are establishing an independent Pipecat Foundation together with many organisations that contribute to and use Pipecat. Kramer wrote that a steering committee will manage it and that it will include companies that compete with Deepgram, according to The Next Web.
3. Model choice. Deepgram says developers will still be able to combine models, tools, transports and APIs from different providers. Stephenson told TechCrunch he wants developers using "the best models for your use case," even when those are not Deepgram's.

A permissive licence such as BSD means the current code can be used and forked regardless of what the owner does later. That is the practical safety net: even in the worst case, the code released so far cannot be taken back. What a licence cannot guarantee is where the maintainers spend their time, which integrations get priority, or how the hosted services are priced. The foundation, if it is truly independent, is meant to cover that gap, but the details of its governance had not been published in the sources we read.

The Real-Time Context Standard

The second announcement is technical. Deepgram says today's voice systems pass words from one model to the next but lose the audio around them: tone, who is speaking, what is happening in the background. RCS, co-developed with the Pipecat team, adds three kinds of metadata:

• Context from the audio, such as a whisper or an unclear word.
• Instructions for how a language model should use those signals.
• Guidance for how the response should sound.

Deepgram's examples are an agent that asks for clarification when it did not hear clearly, or one that answers quietly when someone whispers. Adapters translate the signals and delivery instructions for each model, so the same approach works across providers. The specification is open for testing and feedback, and Deepgram points to a demo repository under the pipecat-ai organisation on GitHub.

Whether RCS becomes a real standard depends on whether other model providers implement it. At announcement, it is a protocol proposed by Deepgram and the Pipecat team; we saw no statement of adoption from other vendors.

What it means

For Pipecat users. Kramer wrote that "for Daily's customers and the Pipecat ecosystem, nothing will change," as quoted by The Next Web. In the short term that is plausible: the code and the licence do not change with the deal. The things to watch are the roadmap and the default integrations. A framework owned by a speech vendor may, over time, make the vendor's models the smoothest path.

For competitors. Voice agent infrastructure companies such as LiveKit and Vapi now compete with a bundle that includes a popular open-source framework, a transport network and speech models under one owner. That bundling is the strategic point of the deal.

For the industry. Open-source orchestration layers have become valuable enough to buy. The pattern is familiar from other parts of the AI stack: the open framework brings developers, and the owner sells hosting, models and support around it. The foundation structure is an attempt to keep the developer trust that made Pipecat popular.

What we don't know yet

• The price and the deal's closing conditions; the announcement describes Deepgram as "acquiring" Daily, and we saw no statement on closing.
• How the Pipecat Foundation will be governed, who sits on the steering committee, and who owns the Pipecat trademark.
• What happens to Daily's hosted products' pricing and to customers who also use competing speech vendors.
• Whether other model makers will adopt RCS.
• Independent performance numbers for the new Deepgram models and Global API; the figures so far are Deepgram's.

What to watch next

Look for the foundation's charter and steering committee list, the first Pipecat releases after the deal, whether RCS support appears in other vendors' SDKs, and the public launch details of the Deepgram Voice Agent Platform.

Sources

• Deepgram, "Deepgram is acquiring Daily, makers of Pipecat," October 29, 2026
• TechCrunch, "Deepgram acquires Pipecat-maker Daily to bolster its voice AI infrastructure," October 29, 2026
• The Next Web, "Deepgram buys Pipecat maker Daily and keeps the voice AI framework open"
• Pipecat