ByteDance Speech Synthesis Model 2.0 (Doubao TTS 2.0) Guide

What is it?

Doubao TTS 2.0 is a high-quality Chinese voice synthesis service from ByteDance (the company behind TikTok), built on large language model technology. It produces natural and fluid voice output.

Important: ByteDance has many similarly-named speech synthesis products. This guide specifically covers the "Speech Synthesis Model → Large Model Speech Synthesis API" — make sure you select the right one, otherwise it won't work.

How to Get Started

Step 1: Register and Log In

1. Visit Volcengine Console
2. Register an account and log in
3. Complete identity verification (required)

Step 2: Create an Application

1. In the console, find Doubao Speech Synthesis Model 2.0

Or go directly to: https://console.volcengine.com/speech/app

2. Note: Switch to "Legacy" mode in the upper left corner — the new interface is cluttered and hard to navigate

3. Navigate through: Legacy → Application Center → Application Management → Create Application

4. Fill in an application name and description (any value works)
5. Important: You must check the "Doubao Speech Synthesis Model 2.0 → Doubao Speech Synthesis Model 2.0 Character Version" checkbox

Step 3: Activate Free Trial or Paid Plan

1. Visit the speech synthesis page: https://console.volcengine.com/speech/service/10035
2. At the top, select the application you just created
3. Click "Free Trial" to activate

Step 4: Copy Configuration Details

Scroll down on the page and copy the following:

1. App ID: Visible at the bottom of the page
2. Access Token: To the right of the App ID — click to copy

Using It in pyVideoTrans

Step 1: Enter Configuration

1. Open the software and go to Menu → TTS Settings → Doubao Speech Synthesis Model 2.0
2. Enter your App ID and Access Token
3. Click the test button to verify the connection
4. Save after a successful test

Step 2: Select a Voice

1. Return to the main interface
2. Select Doubao Speech Synthesis Model 2.0 from the Voice Channel dropdown
3. Choose the target language
4. Select a voice from the Voice dropdown — click to preview

Important Notes

• Many voices require separate activation and purchase in the ByteDance console
• If a voice fails or is unavailable, you likely need to purchase it separately
• Make sure you've selected "Large Model Speech Synthesis API" and not a similar product
• The App ID and Access Token must correspond to the application you created

Troubleshooting

Issue	Solution
Can't find Doubao Speech Synthesis Model 2.0	Switch to "Legacy" mode and follow the navigation path
Test failed	Verify your App ID and Access Token are correct
Some voices unavailable	You need to separately activate and purchase those voices in the ByteDance console
Identity verification prompt	Complete identity verification in the Volcengine console