> ## Documentation Index
> Fetch the complete documentation index at: https://docs.unpod.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Pipecat

> Run the Unpod brain inside an open-source Pipecat voice pipeline - drop-in FrameProcessor, fully self-hosted.

Pipecat is an **open-source voice pipeline** you host yourself. Unlike the SIP-trunk
providers, you don't configure Pipecat as a carrier - instead you drop the Unpod
brain **into** a Pipecat pipeline as a `FrameProcessor`. Audio that Pipecat captures
flows through your `DialogMachine`, and the reply streams back out the same pipeline.

<Note>
  Pipecat is a **runtime**, not a SIP provider. There's no Dashboard/API provider
  config - you wire it in code. The full, working integration lives in the embedding guide.
</Note>

<Card title="Pipecat embedding guide" icon="mic" href="/superdialog/embedding-guides/pipecat">
  The drop-in `FrameProcessor`, install steps, and a complete pipeline example.
</Card>

## How it fits

<div className="uf-flow">
  <div className="uf-node">
    <span className="uf-ic">🎙️</span>
    <span className="uf-t">Mic / SIP audio</span>
    <span className="uf-s">Any Pipecat transport</span>
  </div>

  <div className="uf-arrow"><span className="uf-lbl">① frames</span></div>

  <div className="uf-node is-accent">
    <span className="uf-ic">⚙️</span>
    <span className="uf-t">Pipecat pipeline</span>
    <span className="uf-s">STT · VAD · TTS stages</span>
  </div>

  <div className="uf-arrow is-loop">
    <span className="uf-line fwd"><span className="uf-lbl">② user turn</span></span>
    <span className="uf-line bak"><span className="uf-lbl">③ reply tokens</span></span>
  </div>

  <div className="uf-node is-brain">
    <span className="uf-ic">🧠</span>
    <span className="uf-t">Unpod FrameProcessor</span>
    <span className="uf-s">DialogMachine</span>
  </div>

  <div className="uf-arrow"><span className="uf-lbl">④ TTS stream</span></div>

  <div className="uf-node">
    <span className="uf-ic">🔊</span>
    <span className="uf-t">Audio out</span>
    <span className="uf-s">Low-latency reply</span>
  </div>
</div>

* **Bring audio in** - any Pipecat transport (WebRTC, Daily, telephony) feeds frames.
* **Insert the brain** - add the Unpod `FrameProcessor` so each user turn drives your `DialogMachine`.
* **Stream the reply** - tokens stream back through Pipecat's TTS stage for low latency.

## When to pick Pipecat

| Choose Pipecat when…                                  | Otherwise consider               |
| ----------------------------------------------------- | -------------------------------- |
| You want a fully self-hosted, open-source pipeline    | **Vapi** (managed)               |
| You already run a Pipecat app and want to add a brain | **Unpod Voice** (native runtime) |
| You need fine control over each pipeline stage        | **LiveKit** (low-latency SIP)    |

<CardGroup cols={2}>
  <Card title="Bring your agent" icon="puzzle" href="/speech-stack/bring-your-agent">
    The adapter pattern any runtime uses.
  </Card>

  <Card title="All platforms" icon="blocks" href="/telephony/integrations/overview">
    Back to the integrations overview.
  </Card>
</CardGroup>