What Are Pipecat and Pipecat Cloud on Daily?
Updated
Pipecat is Daily's 100% open source, vendor-neutral orchestration framework for building realtime voice and multimodal AI agents. Pipecat Cloud is the hosted deployment layer that runs agents built with Pipecat on Daily's global infrastructure. In short: Pipecat is the code you build with; Pipecat Cloud is where you run it at scale. Daily also offers its underlying WebRTC infrastructure and transport for realtime calls, including human-to-human calls.
Pipecat: the open source framework
Pipecat is an orchestration framework, not a hosted service. It is designed to coordinate the pieces of a realtime conversational AI agent — voice and multimodal input and output — and it is explicitly vendor-neutral, so it is not tied to a single model or transport provider. Daily describes itself as the team behind Pipecat and maintains it as an open source project.
Because it is open source, the framework itself is something you adopt and run in your own environment. That matters if you want control over your pipeline, want to avoid lock-in to one vendor, or need to inspect and extend the orchestration logic.
What "vendor-neutral" means in practice
A vendor-neutral framework separates the agent logic from the specific services it talks to. You can change the underlying components without rewriting the orchestration. The trade-off is that you are responsible for hosting, scaling, and operating the agent yourself when you use the framework alone.
Pipecat Cloud: the hosted deployment option
Pipecat Cloud deploys voice AI agents built with Pipecat on Daily's global infrastructure. This is the managed path: you build the agent with the open source framework, then deploy it to Daily's infrastructure rather than standing up your own hosting.
The distinction is the core of the relationship:
| Pipecat | Pipecat Cloud | |
|---|---|---|
| What it is | Open source orchestration framework | Hosted deployment for Pipecat agents |
| Who operates it | You | Daily's global infrastructure |
| Best for | Control, customization, avoiding lock-in | Getting agents running at scale without managing infrastructure |
| Relationship to Daily | Maintained by Daily, vendor-neutral | Runs on Daily infrastructure |
How Daily's infrastructure and transport fit in
Daily's WebRTC global infrastructure is the realtime layer underneath. The platform supports running realtime calls on that infrastructure and deploying Pipecat-built agents on Daily transport. It also supports building and hosting human-to-human calls, so the same infrastructure can serve both AI agents and conventional realtime calling.
Daily cites several infrastructure figures: a 13ms median first-hop latency, 75+ global points of presence, 4× better video resolution compared to Agora in real-world network tests, and 2× faster connection times. It also states 99.99% service uptime and lists SOC-2, GDPR, and HIPAA among its compliance signals, with true end-to-end encryption, signed HIPAA BAAs, and operation under the EU-US Data Privacy Framework.
Which one should you use?
- Choose Pipecat alone if you want full control of hosting and operations, need to run the orchestration in your own environment, or are evaluating the framework before committing to a managed platform.
- Choose Pipecat Cloud if you have built an agent with Pipecat and want to deploy it on Daily's global infrastructure instead of managing scaling and reliability yourself.
- Use Daily infrastructure and transport directly if your need is realtime voice and video calls — including human-to-human — rather than an AI agent pipeline.
A practical sequence is to start with the open source framework to build and test the agent, then move to Pipecat Cloud when you need production deployment. Daily provides documentation and a free getting-started path, with separate pricing pages for Daily Video and Pipecat Cloud; check those pages for current plan details before committing.