Pipecat Integration

Pipecat is an open-source framework for building real-time voice and multimodal AI agents. The Pinch Pipecat plugin allows developers to add real-time voice translation directly into Pipecat pipelines.

Prerequisites

Before using the Pinch Pipecat plugin, ensure you have:

Python 3.10+
A Pinch API key
Pipecat installed

You can generate an API key from the Pinch Developer Portal >

Installation

Install the Pinch Pipecat plugin (installs the required dependencies):

pip install pipecat-plugins-pinch

If you plan to run the Daily transport example:

pip install "pipecat-ai[daily]" python-dotenv

Usage

The following example adds Pinch translation to a Pipecat pipeline.

from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner

from pipecat.plugins.pinch import PinchTranslatorService, TranslatorOptions

translator = PinchTranslatorService(
    options=TranslatorOptions(
        source_language="en-US",
        target_language="es-ES"
    )
)

pipeline = Pipeline([
    translator
])

runner = PipelineRunner()
runner.run(pipeline)

This pipeline receives speech in English and outputs translated speech in Spanish.

Running the Example

The git repository includes a working example that demonstrates real-time translation using the Daily transport.

Install required dependencies:

pip install "pipecat-ai[daily]" python-dotenv

Run the example:

python examples/daily_basic.py

Translator Configuration

The TranslatorOptions class allows you to configure translation behavior.

Option	Description
`source_language`	Language of the incoming speech
`target_language`	Language to translate into
`voice_type`	Output voice mode (e.g. cloned voice)

Example configuration:

TranslatorOptions(
    source_language="en-US",
    target_language="fr-FR",
    voice_type="clone"
)