Overview

Now you can handle everything in one place. This guide shows you how to create and manage evaluations directly from your workspace, without any need for the SDK or file uploads.

You can evaluate the naturalness or quality of individual audio files.
1

Log in to Workspace

2

Create a New Evaluation

Go to the Evaluations page and click + New evaluation.

3

Make your selection

Choose how you’d like to create your new evaluation.

4

Choose Models

Add models you’d like to evaluate.

Select the provider of the models you’d like to evaluate.

Select the model you’d like to evaluate.

Select the voice you’d like to evaluate.

5

Choose Scripts

Select the scripts you’d like to use for your evaluation.


6

Set up Instructions

Choose an instruction type (Do, Warning, or Don’t) and provide clear guidance to help participants complete the evaluation effectively.

You can also include detailed examples for clarity.

7

Set up Questions

Now, you can create questions for your evaluation. You can use pre-built templates or create your own questions. Click Add question to start customization.

Once added, you can further refine your custom questions:

8

Set up voting and assignment settings

Fill in the basic information of a new evaluation, such as the title and a detailed description.

9

Review generated audio files

You can review the generated audio files on the Preview page.

10

Review evaluation details

Review all the details of your evaluation on the summary page. Ensure that all the information is correct before confirmation.

If necessary, proceed to payment section to confirm your evaluation.

11

Request Complete and View Evaluation Results

Once you finish the final review, your evaluation request is confirmed.

You can view the details of your completed evaluation from the Workspace or through email.

Once your evaluation completes, a report will be automatically generated and you will be notified via email.

These advanced features allow you to:

  • Register your own voice synthesis models via API
  • Upload your own custom scripts for evaluation
  • Easily compare your models with other voice synthesis models