How to Use Index TTS Online

Open the online tool on the homepage. The left side contains your inputs and the Generate Speech button. The right side initially shows Audio Examples, with playable Index TTS 2.5 and Index TTS 2 samples. On a narrow screen, these areas stack vertically, so scroll below the inputs to find the examples or your result.

Sign in before uploading audio or generating speech. If you are signed out, starting one of those actions opens the sign-in window. Complete the prompts, then return to the tool and check your inputs. Use the same account later when you want to find a recording in My Creations.

The sample players are useful for hearing example speech before you begin. Their model names describe those samples; they are not model-selection buttons. For this walkthrough, keep the task simple: prepare one short script and one authorized reference recording, then make a first test before working on a longer project.

Click the Text field and type or paste your script. The counter beside the label shows how much of the 1,000-character allowance you have used. For a first attempt, one or two complete sentences are enough to check the voice, pronunciation, and pauses without committing a full passage.

The screenshot uses this practice text: Welcome to our guide. Today, we will create a short voice recording with Index TTS. You can use the same text or substitute a short introduction for your own project. This is an input example, not a claim about a tested generation result.

Read the field after pasting, especially the final sentence. Remove headings, editing notes, and instructions that should not be spoken. Use punctuation to separate ideas. If your script is longer than the limit, split it at natural sentence boundaries and keep the complete original in your own document for reference.

Under Reference voice · Required, click Upload reference audio and choose a recording from your device. The interface recommends a clean 3–10 second clip and lists WAV, MP3, and M4A. Use a voice you own or have permission to use for the intended recording.

Listen to the file before uploading it. Choose one clearly audible speaker, with little background noise, no competing music, and a steady recording level. A short, clear sample is more useful for your first comparison than a noisy conversation with several voices. Keep a copy of the reference so you can reuse it for later sections.

Wait for the upload to finish and confirm the selected filename in the upload area. If it reports a failure, retry or choose another playable file. Selecting a file does not by itself confirm that the upload succeeded. The required reference establishes the speaker input; an optional emotional recording in Advanced settings does not replace it.

Click Advanced settings to expand the additional controls. You can leave the initial values unchanged for a first test. Collapsing this section only hides the controls; it does not turn the settings off. If you change a slider and then close the section, that selection still applies to your next request.

Emotional audio is an optional reference for delivery style. Use Upload emotional audio when you have a suitable, authorized recording and want to explore its emotional influence. Keep the required speaker reference in place. Wait for this extra upload to finish before generating, and check its displayed filename just as you would for the main reference.

Strength controls emotional style-transfer strength, on a scale from 0 to 1. The initial value is 0.8. It is not a playback-speed slider. Begin near the existing value and compare a small adjustment before making a large change. The question-mark controls provide short explanations of the fields.

Emotional strengths provides separate sliders for Happy, Angry, Sad, Afraid, Disgusted, Melancholic, Surprised, and Calm. Each runs from 0 to 1 in 0.1 steps. The screenshot shows the initial mix, including nonzero Sad, Afraid, Melancholic, Surprised, and Calm values. Record the values you use and change one at a time so that comparisons remain meaningful.

Check the text, the required reference filename, and any advanced settings. Then click Generate Speech. If you have not signed in, complete sign-in when prompted. If an input needs attention, read the error message and correct that issue before trying again.

During generation, the tool shows a synthesis status and progress display. Allow the request to finish rather than submitting the same text repeatedly. A percentage is not an exact time estimate, and reaching the end of the initial progress can still be followed by a wait for the playable audio to become available.

Watch the output area for the completed recording. If it still says the audio is being generated, wait or check the task in My Creations. When a request fails, retain your script and read the message before changing several inputs. For account or balance issues, check your signed-in account and available credits before making another attempt.

After a successful generation, the result area shows Your Generated Speech with an audio player. Press Play and follow your script while listening. Check that the entire passage is present, including the final sentence. Pay attention to names, numbers, pauses, and whether the delivery fits your purpose.

Use Download below your generated player to save the audio. Open the downloaded file on your device to verify that it is the recording you intended to keep. The screenshot here shows the output area's initial sample state, so do not mistake an official example for your completed request.

To compare a revision, save your preferred version first. Edit the text, replace a reference, or adjust an emotion setting, then use Generate Again. This submits another generation with the current inputs rather than replaying the existing recording. Change one variable per comparison and listen to the same sentence in both results.