> For the documentation index, fetch https://sync.so/docs/llms.txt. Append .md to a page URL for Markdown. Documentation-search MCP: https://sync.so/docs/_mcp/server.

# Edit Dialogue — Studio

> Change or remove words in a video without re-recording, in the speaker's own voice.

**edit dialogue** changes what someone says in a video without re-recording it. You edit the words in a transcript, Sync re-speaks only the edited words in the speaker's own voice, and then lip-syncs only those parts of the video. Everything you did not edit keeps the original recording.

> **Info**
>
> edit dialogue is in **beta**. It works best on a video with one person speaking clearly to camera for at least 10 seconds. To do the same through the API, see [Transcription and Dialogue Editing](/docs/developer-guides/dialogue-editing).

## When to use edit dialogue

| You want to                                                                                     | Use                                                                                                            |
| :---------------------------------------------------------------------------------------------- | :------------------------------------------------------------------------------------------------------------- |
| Fix a wrong word, a name, a number, or a short phrase, and keep the original voice and delivery | **edit dialogue**                                                                                              |
| Remove a few words, a filler, or a false start                                                  | **edit dialogue**                                                                                              |
| Replace the whole script, or most of a line                                                     | **lipsync video** with new audio: upload or record it, or use [text to speech](/docs/web-app-guides/text-to-speech) |
| Use a different voice than the speaker's                                                        | **lipsync video** with [text to speech](/docs/web-app-guides/text-to-speech) and the voice you choose               |
| Put the video into another language                                                             | [**translate & dub**](/docs/web-app-guides/dubbing-ui)                                                              |

Each changed word can become at most a few words, so edit dialogue fixes lines rather than rewriting them. If your change needs more room than that, re-voice the line instead.

## Edit a video's dialogue

#### Open edit dialogue

On **home**, select the **edit dialogue** card. You can also select **edit dialogue** on a finished result to continue from that video.

#### Upload the video

Drop in or select your video. Transcription starts on its own, usually before the upload finishes. Until it is ready, Studio shows **the transcript starts once the video finishes processing**.

#### Change or remove words in the transcript

In the **transcript** column, click a word to select it.

* To **change** it, double-click the word or start typing, then enter the new wording.
* To **remove** it, press Backspace. Removed words are shown struck through. Click a removed word to restore it.

Use **undo**, **redo**, or **discard edits** to step back. **discard edits** returns to the original dialogue.

#### Preview the audio

Select **preview audio**. The button shows how many edits are waiting, for example **preview audio (2)**. Sync generates the new audio, and you can switch between the **original** and **edited** takes to compare them. The footer shows **est. duration** and **est. cost** for the synced video.

#### Sync

When the preview sounds right, select **Sync**. Sync lip-syncs the edited parts with the sync-3 model and adds a new video to your history, labelled **changed dialogue at** followed by the time. The editor stays open, so you can keep editing and sync again.

Each **Sync** creates a new video, so preview your edits first and sync when you are ready. If you open a video you made with edit dialogue again, the original transcript and edits come back without transcribing it again.

## Which voice is used

edit dialogue always uses **the speaker's own voice**. Sync makes an automatic clone of the speaker from the video's audio and uses it only for the edited words.

* There is no voice picker. You cannot choose a library voice or one of your custom voices for edit dialogue.
* You do not need an ElevenLabs account or API key.
* This automatic clone is not added to your voices and does not use one of your plan's voice clone slots.
* Background music and sound effects in the original audio are kept under the edited speech.

## What you can and can't change

| Limit             | Value                                                                                                           |
| :---------------- | :-------------------------------------------------------------------------------------------------------------- |
| Speakers          | One speaker per video. Videos with several people talking are not supported yet.                                |
| Changing a word   | Up to **3 words** in place of one word, and about two words' worth of extra characters                          |
| Adding words      | No free insertion between words. Grow a neighbouring word into up to 3 words instead.                           |
| Removing words    | About **3 seconds** per section of speech. Removing more in one place stops the video from matching the speech. |
| Preview window    | **10 minutes** of video per preview. On a longer video, keep each batch of edits within one 10-minute stretch.  |
| Edits per preview | Up to 100                                                                                                       |
| Video length      | Videos up to 60 minutes can be transcribed                                                                      |
| Languages         | 29 languages                                                                                                    |
| Plan limit        | The total length of the edited parts must fit your plan's maximum output length                                 |
| Not supported     | Captions, more than one speaker, and fine control over pacing and emotion                                       |

The new words are spoken where the original words were. Sync can slow the picture down slightly to make room, but a much longer wording has nowhere to go. If Studio warns that an edit **may be too long to fit**, try a shorter wording before you preview.

## What it costs

You pay for the synced video as a [sync-3](/docs/models/sync-3) generation.

* Usually you're billed for the audio around your edits that gets lip-synced, up to the preview's length. This can include nearby speech, not just the words you changed.
* If the edits can't be lip-synced section by section, the sync may instead cover the whole preview and be billed for its full length. This fallback only applies to a preview that can still be submitted.

After a usable preview of your current edits finishes, **est. cost** shows the estimated price. A preview that cannot be submitted has no estimate and cannot be synced. The estimate may read **free** when the sync is covered by your free allowance.

Every **Sync** is a separate generation and is billed separately. See [Billing](/docs/product/billing) for model rates.

## Common messages

| Message                                                                         | What it means                                                           | What to do                                                                       |
| :------------------------------------------------------------------------------ | :---------------------------------------------------------------------- | :------------------------------------------------------------------------------- |
| **Change a word to sync**                                                       | You have not edited the transcript yet.                                 | Change or remove at least one word, then select **preview audio**.               |
| **preview the audio first**                                                     | Sync needs a finished preview of your current edits.                    | Select **preview audio** and wait for it to finish.                              |
| **That edit is too long to fit** or **"…" is too long to fit**                  | The new wording needs more time than the words it replaces.             | Use a shorter wording, then preview again.                                       |
| **up to 3 words in place of one** or **keep the change within about two words** | The replacement is longer than one word can take.                       | Shorten it, or split the change across neighbouring words.                       |
| **too much removed in one place**                                               | Too much speech is removed from one section to keep the video in sync.  | Restore some words there, or spread the removals out. Your other edits are kept. |
| **edits span more than 10 minutes**                                             | One preview covers up to 10 minutes of video.                           | Keep the batch within one 10-minute stretch, then preview again.                 |
| **this video has N speakers** or **This video has more than one speaker**       | More than one person speaks in the video.                               | Use a video with a single speaker.                                               |
| **The voice sample is too short**                                               | There is not enough clear speech to clone the voice.                    | Use a longer clip with clear speech, then preview again.                         |
| **This speaker's voice can't be cloned**                                        | The voice provider does not allow cloning this voice.                   | Edit a different video or speaker.                                               |
| **This voice needs verification**                                               | The voice provider requires verification before this voice can be used. | Contact support.                                                                 |
| **These edits are longer than your plan allows**                                | The edited parts are longer than your plan's output limit.              | Undo some edits, or upgrade your plan.                                           |
| **Dialogue editing supports videos up to 10 minutes**                           | The stretch you are editing is longer than one preview can cover.       | Trim the video, or edit a shorter stretch.                                       |
| **this video has no audio** or **no speech in this video**                      | There is no dialogue to transcribe.                                     | Use a video with speech, or use **lipsync video** with a voiceover.              |
| **Too many previews right now**                                                 | You requested many previews in a short time.                            | Wait a moment and try again.                                                     |

If a preview fails for another reason, try again. If it keeps failing, contact [support@sync.so](mailto:support@sync.so).

## Frequently Asked Questions

#### Do I need to re-record if I said the wrong word?

No. Use edit dialogue to change or remove the word. Re-record only when you need to change most of a line or the whole script.

#### Can I pick another voice for the new words?

No. edit dialogue always clones the speaker in the video. To use a different voice, create the audio with [text to speech](/docs/web-app-guides/text-to-speech) and lip-sync it with **lipsync video**.

#### Does the video keep its length?

Mostly. A changed word is spoken in the space of the original word, and the picture is slowed slightly only when the new wording needs a little more time. Removing words shortens the video.

#### Is edit dialogue available in the Adobe Premiere or DaVinci Resolve plugins?

No. Use it in Studio, through the [API](/docs/developer-guides/dialogue-editing), in the [ChatGPT app](/docs/plugins-and-extensions/chatgpt-plugin), or with the [MCP server](/docs/plugins-and-extensions/mcp-server).