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

# Fit it into your stack

> See where speech to speech sits in your system, and how a caller's audio reaches the agent: from a phone line, a web page, LiveKit or your own code.

## The parts of a call

<Flow caption="One call: your server holds one WebSocket per call">
  <FlowNode title="Caller" icon="phone">On a phone, a web page or an app</FlowNode>
  <FlowLink label="audio" both />
  <FlowNode title="Your server" icon="server">Opens the session and passes audio both ways</FlowNode>
  <FlowLink label="WebSocket: audio, text, events" both />
  <FlowNode title="Fluid 1" icon="waveform-lines" highlight>Hears the caller, answers, speaks</FlowNode>
</Flow>

<Flow caption="When a turn needs your systems, the agent delegates it to the back office">
  <FlowNode title="Fluid 1" icon="waveform-lines" highlight>Holds the conversation</FlowNode>
  <FlowLink label="delegation, then the result" both />
  <FlowNode title="Back office" icon="database">Your backend, or a model KugelAudio runs for you</FlowNode>
</Flow>

| Part | What it does | Who runs it |
|---|---|---|
| The caller's connection | Carries the caller's audio: a phone line, a web page or an app | You, or your phone provider |
| Your server | Starts the session, streams the caller's audio in and the agent's audio out | You |
| Fluid 1 | Listens, decides when to answer, speaks, and delegates what it cannot answer | KugelAudio |
| Back office | Does the work a turn needs and sends the result | [You](/speech-to-speech/your-backend), or [KugelAudio](/speech-to-speech/back-office) |

The delegation travels on the same WebSocket as the audio, so a call needs one connection, not
two.

## Keep your key on your server

The WebSocket takes your API key. Open it from your server, never from a web page or an app:
anyone who loads the page can read a key that is in it. The caller's device talks to your
server, and your server talks to KugelAudio.

## Pick how the audio reaches the agent

| The caller is on | How the audio gets there | Audio format |
|---|---|---|
| A phone line (Twilio, Telnyx, a SIP gateway) | Your phone provider streams the call to your server, which passes each frame on | `audio/pcmu` or `audio/pcma` at 8 kHz, both ways, unchanged |
| A web page or an app | The page or app streams its microphone to your server, which passes it on | `audio/pcm` at 24 or 16 kHz |
| LiveKit Agents | LiveKit's OpenAI plugin holds the WebSocket for you | Chosen by the plugin |
| Your own microphone, to try it | A script streams the microphone and plays the reply: the [quickstart](/speech-to-speech/quickstart) | `audio/pcm` at 24 kHz |

### A phone line

<Flow>
  <FlowNode title="Caller" icon="phone" />
  <FlowLink label="phone network" />
  <FlowNode title="Phone provider" icon="tower-broadcast">Twilio, Telnyx or a SIP gateway</FlowNode>
  <FlowLink label="media stream, G.711" both />
  <FlowNode title="Your server" icon="server" />
  <FlowLink label="WebSocket, the same G.711" both />
  <FlowNode title="Fluid 1" icon="waveform-lines" highlight />
</Flow>

Phone systems send 8 kHz G.711, and the agent answers in the same format, so your server
forwards frames without converting them. Ask for the format when you start the session:
[Connect a phone line](/speech-to-speech/during-a-call#connect-a-phone-line).

### A web page or an app

<Flow>
  <FlowNode title="Web page or app" icon="globe">Microphone and speaker</FlowNode>
  <FlowLink label="your own connection" both />
  <FlowNode title="Your server" icon="server">Checks who the user is</FlowNode>
  <FlowLink label="WebSocket" both />
  <FlowNode title="Fluid 1" icon="waveform-lines" highlight />
</Flow>

Your page sends the microphone to your server over a connection you already have, a WebSocket
for example, and plays what comes back. Your server decides who may talk to the agent, so your
own sign-in and rate limits apply.

### LiveKit Agents

LiveKit's OpenAI plugin connects to speech to speech in place of your own WebSocket code. See
[speech to speech in LiveKit](/integrations/livekit#speech-to-speech-with-fluid-1).

## Build it to last

- **One session per call.** A session is one conversation on one socket, lasts at most an hour
  and is not stored. Open a new one for every call. See [limits](/speech-to-speech/limits).
- **Start the session before the caller speaks.** The model, the instructions and the audio
  format are fixed at the start. You can [add to the instructions](/speech-to-speech/your-backend#change-how-the-agent-behaves)
  during the call.
- **Stream audio as it arrives.** Send the caller's audio in small frames as you receive it,
  and play the agent's audio as it comes. Do not wait for a whole sentence.
- **Stop playback when the caller interrupts.** The agent stops talking when the caller cuts in.
  If you buffer its audio, [drop the queue](/speech-to-speech/during-a-call#stop-playback-when-the-caller-interrupts).
- **Answer delegations quickly.** The caller waits for your result. For slow work,
  [report progress](/speech-to-speech/your-backend#report-progress-on-slow-work) so the agent can tell them.
- **Plan for a dropped connection.** A broken socket ends the session. Start a new one with what
  you kept, and the agent [carries on](/speech-to-speech/during-a-call#continue-after-a-session-ended).
- **Retry a refused session with backoff.** See [retry a refused session](/speech-to-speech/limits#retry-a-refused-session).

## Next steps

<CardGroup cols={2}>
  <Card title="Talk to the agent" icon="microphone" href="/speech-to-speech/quickstart">
    Talk to it from your microphone in a few minutes
  </Card>
  <Card title="Answer from your backend" icon="server" href="/speech-to-speech/your-backend">
    Look things up and have the agent tell the caller
  </Card>
</CardGroup>