Skip to navigation

Text-to-Speech and Voices — Studio

Text to speech turns a script into audio in the voice you choose, and Studio then lip-syncs your video or image to it. You can use a ready-made voice or clone your own. You don’t need an ElevenLabs account or API key. For the API, see Text-to-Speech and Voice Cloning.

To change a few words a person already says in a video, in their own voice, use edit dialogue instead.

Create speech from text

1

Start a lipsync video

On home, select lipsync video and add your video or image.

2

Choose text to speech as the audio

In the audio input, select Text to speech.

3

Pick a voice

Under Voice, open the voice picker and choose a voice. Use Find voice to search by name.

4

Write the script

Type your script in the Text to speech box. Scripts can be up to 5,000 characters.

5

Generate and listen

Select Generate and listen to the audio. If you change the script or voice, select Regenerate.

6

Sync

When the audio sounds right, select Sync to create the lip-synced video.

Audio tags

Add audio tags in square brackets to direct the delivery, for example [happy], [whispers], or [laughs].

[excited] We just launched something new. [whispers] And you're the first to hear about it.

Voice settings

Open the voice settings to adjust two values:

SettingDefaultEffect
Stability0.50Lower values sound more expressive and varied. Higher values sound steadier and more consistent.
Similarity Boost0.75Higher values follow the original voice more closely. Lower values allow more variation.

Limits and cost

Text to speech is billed per second of generated audio, in addition to the lip-sync generation. See Billing for rates, or Credit Billing for credit plans.

On the free plan:

  • With sync. labs’ built-in ElevenLabs access, each request is limited to 20 seconds of generated speech and 320 input characters.
  • With your own connected ElevenLabs API key, the limit is 20 estimated seconds per request, without the 320-character input ceiling. The monthly free-generation limits still apply.
  • Text to speech and dubbing share 10 audio generations a month. Up to 3 of them can be dubbing.

With built-in access, Studio shows Free tier text-to-speech is limited to 20 seconds and 320 input characters when the input exceeds the limit. With your own key, the duration message is Free tier text-to-speech is limited to 20 estimated seconds. Studio also tells you when you have used all your free generations for the month. Shorten the script, wait for the next month, or upgrade your plan.

Choose a voice

The voice picker shows two groups:

  • Cloned voices (N): the custom voices your organization has cloned.
  • Eleven Labs (N): ready-made ElevenLabs voices.

If you connect your own ElevenLabs API key, the picker shows every voice that key can access.

Some voices can’t be used. A voice marked Blocked by ElevenLabs. Choose another voice. or Verification required. Choose another voice. is restricted by ElevenLabs, so pick a different voice.

Clone a voice

Cloning creates a custom voice from a recording, so you can generate new speech that sounds like that person. Clone only voices you have permission to use.

1

Open the clone form

In the voice picker, select Clone voice. The Create custom voice form opens.

2

Name the voice

Enter a Voice name. Names can be up to 100 characters and must be unique in your organization.

3

Add a sample

Under Upload or record audio sample, upload an audio file, record one, or extract the audio from a Video or Audio file.

4

Create the voice

Select Create voice. The new voice appears under Cloned voices in the picker.

Sample guidelines

  • Aim for 1 to 5 minutes of continuous, clear speech from one person, without music or background noise.
  • Each sample must be at least 7 seconds and at most 5 minutes long. You can add up to 10 samples.
  • Each file can be up to 10 MB, or up to 50 MB when your own ElevenLabs API key is connected.
  • When you extract from a video, only the first 2 minutes of its audio are used.

Clone slots per plan

Each plan can keep a set number of cloned voices at once:

Usage-billed planCredit planCloned voices
FreeFree1
HobbyistHobby3
CreatorStarter, Pro5
GrowthTeam15
Scale—50

Enterprise plans can have a custom allowance. When every slot is in use, Studio shows You have reached your plan’s voice clone limit. Delete a voice you no longer need, or upgrade your plan.

Delete a cloned voice

In the voice picker, delete the voice and confirm Delete “name”?. Deleting removes the voice for everyone in your organization and can’t be undone. It frees a clone slot right away.

Use your own ElevenLabs API key (optional)

You never need an ElevenLabs account to use text to speech, voice cloning, dubbing, or edit dialogue in Studio. Without your own key, Sync’s built-in ElevenLabs access is used and billed through Sync.

If you prefer to use your own ElevenLabs account, open Settings → Integrations, configure the ElevenLabs integration, and paste your API Key. With your own key:

  • ElevenLabs bills the generated audio to your ElevenLabs account. Sync still bills the lip sync.
  • The voice picker shows every voice your key can access.
  • Voice samples can be up to 50 MB each.

If ElevenLabs rejects your key as invalid, Studio disconnects it and goes back to Sync’s built-in access. For key errors, see Integrations.

Frequently Asked Questions

No. Text to speech and voice cloning work in Studio without an ElevenLabs account. Connecting your own key is optional.

edit dialogue always clones the person speaking in the video and does not offer a voice picker. Custom voices are for text to speech.

Delete a cloned voice you no longer need from the voice picker. The slot is available right away.