Deepdub

Text-to-speech service implementation using Deepdub AI’s streaming WebSocket API

Overview

DeepdubTTSService generates speech from text using Deepdub AI’s real-time streaming WebSocket API. It extends Pipecat’s InterruptibleTTSService and pushes TTSAudioRawFrame audio into the pipeline, with support for multiple models (dd-etts-2.5, dd-etts-3.0), voice customization, and accent control.

Source Repository

Source code, examples, and issues for the Deepdub TTS integration

PyPI Package

The pipecat-deepdub-tts package on PyPI

Deepdub AI

Learn more about Deepdub AI and sign up for an API key

Installation

This is a community-maintained package distributed separately from pipecat-ai:

uv add pipecat-deepdub-tts

Prerequisites

Deepdub Account Setup

Before using the Deepdub TTS service, you need:

1. Deepdub Account: Sign up at Deepdub AI
2. API Key: Obtain your API key from your Deepdub account
3. Voice Prompt ID: A voice prompt ID to use for synthesis

Required Environment Variables

• DEEPDUB_API_KEY: Your Deepdub API key for authentication
• DEEPDUB_VOICE_ID: The voice prompt ID used for synthesis

Configuration

api_key · str · required

Deepdub API key for authentication.

voice_id · str · required

Voice prompt ID to use for synthesis.

model · str · default: "dd-etts-2.5"

TTS model name (e.g. "dd-etts-2.5", "dd-etts-3.0").

sample_rate · int · default: "16000"

Audio sample rate in Hz. Valid values: 8000, 16000, 22050, 24000, 44100, 48000.

params · DeepdubTTSService.InputParams · default: "None"

Optional voice customization parameters. See InputParams below.

InputParams

Optional voice customization passed via the params constructor argument using DeepdubTTSService.InputParams(...).

Parameter	Type	Default	Description
locale	str	"en-US"	Language locale for synthesis.
temperature	float	None	Controls output variability.
variance	float	None	Controls variance in generated speech.
tempo	float	None	Speech tempo multiplier.
prompt_boost	bool	None	Enable prompt boosting for improved quality.
accent_base_locale	str	None	Base locale for accent control. All three accent params required together.
accent_locale	str	None	Target accent locale.
accent_ratio	float	None	Accent blending ratio (0.0 to 1.0).

Available parameters and defaults are defined by the integration. See the source repository for the authoritative, up-to-date list.

Usage

import os
from pipecat_deepdub_tts import DeepdubTTSService

tts = DeepdubTTSService(
    api_key=os.getenv("DEEPDUB_API_KEY"),
    voice_id=os.getenv("DEEPDUB_VOICE_ID"),
    model="dd-etts-2.5",
    sample_rate=24000,
    params=DeepdubTTSService.InputParams(
        locale="en-US",
        temperature=0.7,
        tempo=1.1,
        prompt_boost=True,
    ),
)

# ... add tts to your pipeline, e.g. Pipeline([llm, tts, ...])

Compatibility

Tested with Pipecat v0.0.97. Check the source repository for the latest tested version and changelog.