---
title: "How do I make a character speak?"
description: "Add a voice to a finished video take with Add voice, written per scene on ElevenLabs v4; then lay it on the take with the Lip sync tool."
url: https://www.filmsquare.ai/help/models-and-tools/dialogue-and-lip-sync/
lang: en
section: models-and-tools
updated: 2026-11-13
---

Dialogue is made per scene, one voice take per character, and then laid over a video take. There are two steps: make the voice, then make the mouth follow it with **Lip sync**.

## Make the voice

1. Write the dialogue lines in the scene.
2. Open a finished video take of a shot in that scene and choose **Add voice**. The button shows the price, and it waits until the lines of that shot are known.
3. The voice is spoken on ElevenLabs v4: 2.94 credits · €0.15 per 1,000 characters, about a minute of speech.
4. Pick the dialogue take. The cast member then reads **Dialogue take picked** in the Voices department.

**Add voice** stays off with a reason when the scene has no dialogue lines, when the shot has no lines of its own, when the character has no voice yet ("No voice yet"), or when the shot is not placed in a scene. Dialogue is made per scene, so a shot in no scene cannot have it.

## Make the mouth follow it

1. Open **Create → Tools** and choose **Lip sync**.
2. Choose the video take. It can be up to 30 seconds long and must be at least 512 px high.
3. Under **Voice**, choose **From this shot** and pick a **Dialogue take**, or choose **Upload** and add an audio file. Audio needs at least 2 seconds.
4. Check the price on the button, then make the take. The result is a new take next to the original.

![A shot with its takes, the picked take and a note pinned on the frame](https://www.filmsquare.ai/docs/shots/en/shot-detail-desktop.webp)

If the voice is longer than the clip, the tool shows how the voice fits the clip before you spend anything.

> [!COST]
> The voice and the lip sync are two takes, each with its price shown before you start. A failed run is not charged.

## Good to know

Seedance 2.5 and Kling 3.0 can also generate sound with the video. The Voices department and **Lip sync** are for the case where you want your own lines in a character's voice. See [which model](https://www.filmsquare.ai/help/models-and-tools/which-model/) for what each video model offers.

More in the docs: [Voices](https://www.filmsquare.ai/docs/departments/extra-departments/#voices) and [Tools](https://www.filmsquare.ai/docs/concepts/takes/#tools).
