Overview

The Pipeline is a single, self-configuring component. You hand it the components you need, and based on what you pass in it automatically selects Cascade, Realtime, or Hybrid mode and connects them for you.

Pipeline Code Snippet

A basic Cascade pipeline built from STT, LLM, TTS, VAD, and a Turn Detector.

from zrt import Pipeline
from zrt.plugins import DeepgramSTT, OpenAILLM, ElevenLabsTTS, SileroVAD, TurnDetector
from zrt.inference import AICousticsDenoise

pipeline = Pipeline(
    stt=DeepgramSTT(),
    llm=OpenAILLM(),
    tts=ElevenLabsTTS(),
    vad=SileroVAD(),
    turn_detector=TurnDetector(model="echo-large"),
    denoise=AICousticsDenoise(model_id="quail-vf-2.2-l-16khz"),
)

What’s Next

Plugins

STT

Speech-to-text providers.

TTS

Text-to-speech providers.

LLM

Language model providers.

Realtime

Speech-to-speech realtime models.

Feature

Pipeline Observability

Observe metrics and traces at each pipeline stage.

Modes

Cascade, Realtime, and Hybrid, chosen automatically from your components.

References

Python

Examples

Cascade Basic

Minimal cascade pipeline.

Realtime Basic

Minimal Realtime pipeline.

SDK Reference

Pipeline

Pipeline in the Python API reference.

Avatar Integration Modes

Get started

Modalities

Configure Pipeline

Tools and Capabilities

Turn Detection and Interruptions

Telephony Agents

Context Management

Observability and Monitoring

Pipeline Code Snippet

What’s Next

Plugins

STT

TTS

LLM

Realtime

Feature

Pipeline Observability

Modes

References

Examples

Cascade Basic

Realtime Basic

SDK Reference

Pipeline

​Pipeline Code Snippet

​What’s Next

​Plugins

STT

TTS

LLM

Realtime

​Feature

Pipeline Observability

Modes

​References

​Examples

Cascade Basic

Realtime Basic

​SDK Reference

Pipeline

Pipeline Code Snippet

What’s Next

Plugins

Feature

References

Examples

SDK Reference