> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tryhamsa.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Create a voice over, then turn it into a video

> Use the voice over composer: pick a use case, write a prompt or your script, set dialect, voice, music and files, pick a take, then make a video.

The composer is where you make a voice over: you pick what it is for, write a prompt or paste your script, choose the dialect, voice and music in the bar under the text, and click **Create voice over**. This page walks through each part, then shows how to turn the result into a video.

New to voice overs? Read [Voice overs](https://docs.tryhamsa.com/media/voice-overs/overview.md) first for the difference between **Prompt** and **My script**.

## Before you start

* Credits left on the project owner's account. A voice over costs 10 credits. See [Billing](https://docs.tryhamsa.com/overview/billing.md).
* For **My script**, the voice you want. You can also clone one from the composer.

## Create a voice over

Open **Services** > **Voice overs**. If a voice over is open, click **Create new** first.

![The composer in Prompt mode with Stories selected, a prompt for a winter festival teaser typed in, and the bar under the text with MSA, the Music switch and the paperclip](https://mintcdn.com/hamsa-452fa79a/etUfiFCo8p_6XpY_/images/media/voice-overs/composer-prompt-light.webp?fit=max\&auto=format\&n=etUfiFCo8p_6XpY_\&q=85\&s=8ef6a459ab92e496cade1597dc0ac2a7)

**1. Pick a use case**

In the bar at the top, choose **Stories**, **Commercial**, **Religious** or **Poetry**. Switching use case clears the voice and the track you picked, since each use case has its own voices and music.

**2. Choose how to start**

Pick **Prompt** to describe the idea and let AI write it, or **My script** to have your own text read as it is. Each mode keeps its own text, so you can switch without losing either.

**3. Write**

Type or paste up to 1000 characters. You need at least 20 ("At least 20 characters required"). In **Prompt**, you can instead attach a file and leave the text empty.

**4. Set dialect, voice and music**

Use the bar under the text. What it holds depends on the mode; see [The bar under the text](#the-bar-under-the-text).

**5. Create**

Click **Create voice over**. The page shows "Creating your voice over" and the job appears at the top of the **Job history** with **Processing**.

**6. Pick your take**

In **Prompt**, **Voice over options** opens with three options, each read by its voice. Play them and click **Use** on the one you want. In **My script**, your take is kept for you; there is nothing to pick.

![Voice over options for a Commercial in the UAE dialect with background music: three options read by Hala, Fayez and Nour, each with a Use button and a play button](https://mintcdn.com/hamsa-452fa79a/etUfiFCo8p_6XpY_/images/media/voice-overs/voice-over-options-light.webp?fit=max\&auto=format\&n=etUfiFCo8p_6XpY_\&q=85\&s=d73779abb7ba49ae9b4967e2c0dcaeb5)

Your choice with **Use** is final: a voice over keeps one take, and you can't switch to another option later. To try again with different settings, create a new voice over.

> [!NOTE]
>
> If only one of the three options comes out, it is kept for you as if you had clicked **Use**. Options that failed don't show.

## The bar under the text

| In **Prompt** | What it does |
| - | - |
| Dialect | The Arabic dialect AI writes and reads in. The list shows the dialects that have voices for your use case, and ends with **Project defaults**. |
| **Music** | Off by default. On, AI picks a background track for each option, a different track each. |
| Paperclip (**Attach file**) | Attach one file for AI to work from: JPG, PNG, PDF, DOCX, TXT or MD, up to 35 MB. |

| In **My script** | What it does |
| - | - |
| **Choose voice · required** | Opens **Choose a voice**. You can't create without a voice. Picking a voice also sets the dialect. |
| **Add music** | Opens **Choose background music**. Optional. |

![The composer in My script mode with a news script typed in, the estimate about 20 s, about 10 credits, and the bar under the text showing the voice Rania and the track Morning light](https://mintcdn.com/hamsa-452fa79a/etUfiFCo8p_6XpY_/images/media/voice-overs/composer-my-script-light.webp?fit=max\&auto=format\&n=etUfiFCo8p_6XpY_\&q=85\&s=1b783864f5957f20954f91a91e6c2cae)

The X at the end of the bar folds it away; the settings icon brings it back.

### Choose a voice (My script)

**Choose a voice** has three tabs:

* **All voices**: voices made for the use case you picked.
* **Project voices**: voices cloned in this project.
* **Favorites**: voices you marked as favorites in the [Voices library](https://docs.tryhamsa.com/media/voices/voices-library.md).

Search and filter by language, dialect, style and gender as in the Voices library. Play a voice before you pick it. **Clone voice** opens the clone dialog; the new voice shows under **Project voices**. See [Add a voice](https://docs.tryhamsa.com/media/voices/add-a-voice.md).

![Choose a voice with the All voices, Project voices and Favorites tabs, a search box, Language and Dialect filters, and voices such as Rania, Fayez and Nour, each with Select and a play button](https://mintcdn.com/hamsa-452fa79a/etUfiFCo8p_6XpY_/images/media/voice-overs/choose-voice-light.webp?fit=max\&auto=format\&n=etUfiFCo8p_6XpY_\&q=85\&s=92241905d6f8cdd0730fc5b2986f0bc1)

### Background music

In **My script**, **Choose background music** offers:

| Option | What you get |
| - | - |
| **No music** | "Voice only, no background track." |
| **Auto music** | AI picks a track that fits your script and voice, and how it fades in and out. |
| A track | One of Hamsa's tracks for the use case. Play it first. It fades in at the start and out at the end. |

![Choose background music with No music selected, Auto music next to it, and six tracks such as Morning light and City pulse](https://mintcdn.com/hamsa-452fa79a/etUfiFCo8p_6XpY_/images/media/voice-overs/background-music-light.webp?fit=max\&auto=format\&n=etUfiFCo8p_6XpY_\&q=85\&s=94674e4a582aa3fa9ba407ecf3405b7f)

Tracks you used before in this project come first, under **Recently used**, then **All tracks**. The music always sits well under the voice (about 15% of its volume); you can't change the volume or upload your own track.

### Attach a file (Prompt)

Click the paperclip, then drag a file into **Attach file** or click **Select file**. Supported: JPG, PNG, PDF, DOCX, TXT and MD, up to 35 MB. One file per voice over; **Replace file** and **Remove file** change it.

![The Attach file dialog with Select file and the supported types JPG, PNG, PDF, DOCX, TXT and MD, max size 35 MB](https://mintcdn.com/hamsa-452fa79a/etUfiFCo8p_6XpY_/images/media/voice-overs/attach-file-light.webp?fit=max\&auto=format\&n=etUfiFCo8p_6XpY_\&q=85\&s=24ce25459c78c0cabb562d5b8b4b12fe)

AI treats the file as the main source: it writes from what the file shows or says, and uses your prompt for anything the file doesn't give, like tone, names that must appear, or what to leave out. A product photo with "a 20 second launch ad, warm and upbeat" is a good pair.

## The estimate (My script)

In **My script**, the length and cost show under the text as you type, for example "about 40 s · about 10 credits". Hover it for the word count. It is an estimate from a typical reading pace (about 155 words a minute in English and 115 in Arabic); the voice, numbers and punctuation change the real length. The price doesn't depend on length: "10 credits per creation, whatever the length."

**Prompt** has no estimate, because the script doesn't exist until AI writes it.

## Drafts and text from other pages

* What you type is kept as a draft for this project in this browser ("Saved to drafts" shows when you hover the text). When you come back, **Draft restored** shows with **Undo**, which clears the text.
* Creating a voice over clears the text and the attached file.
* From a transcript or from [Text to speech](https://docs.tryhamsa.com/media/text-to-speech/generate-speech.md), **Use in a voice over** opens the composer in **My script** with that text. Text over 1000 characters asks first, with **Use the first 1000 characters**. If you were already writing a script, **Replace your voice over script?** asks before replacing it; after **Replace it**, "Replaced your earlier script" offers **Undo**.

## Project defaults

To start every new voice over in the same dialect, set **Voice over dialect** in **Defaults for this project**: open it from **Project defaults** at the end of the dialect list, or from Cmd+K (Ctrl+K on Windows). A new, empty composer then starts in that dialect and says "Using your project defaults", with **Change**. Defaults are saved in this browser only. See [Project defaults](https://docs.tryhamsa.com/media/workspace/project-defaults.md).

## If it fails

**Generation failed** says "No credits were deducted for this attempt." Click **Try again** to get the form back with your prompt, music and dialect, and create again.

![A failed voice over open: Generation failed, a note that no credits were deducted, and a Try again button, with the Failed row selected in the Job history](https://mintcdn.com/hamsa-452fa79a/etUfiFCo8p_6XpY_/images/media/voice-overs/generation-failed-light.webp?fit=max\&auto=format\&n=etUfiFCo8p_6XpY_\&q=85\&s=4e456e9c63d72c2ff7cd924017a46f83)

## Turn it into a video

Once you have kept a take, the voice over shows **Turn this into a video**. A video costs 40 credits and is charged only when it finishes.

**1. Open the video settings**

Click **Generate video**. **Video configuration** opens on the **Video** tab.

**2. Start from a preset**

Under **Presets**, pick **Social reel**, **Explainer** or **Story**, then change anything below. A preset you change shows "modified".

**3. Add a reference image or design instructions (optional)**

On **Advanced**, upload a **Reference image** (JPG or PNG, up to 5 MB) for AI to keep in view, and write **Design instructions**, up to 500 characters, for example "Warm cinematic colors, soft natural lighting, minimal modern aesthetic."

![The Advanced tab of Video configuration: Upload an image for the reference image, and design instructions typed in, 71 of 500 characters](https://mintcdn.com/hamsa-452fa79a/etUfiFCo8p_6XpY_/images/media/voice-overs/video-config-advanced-light.webp?fit=max\&auto=format\&n=etUfiFCo8p_6XpY_\&q=85\&s=251619fea2daa5abe80c44d0591e75f1)

**4. Generate**

Click **Generate video**. The video card shows its progress; see [Manage voice overs](https://docs.tryhamsa.com/media/voice-overs/manage-voice-overs.md#videos).

![Video configuration on the Video tab with the Social reel preset selected: Cartoon style, Portrait 9:16, Ken Burns effect and Captions on, Word by word, Caption box on, and Fade as the scene transition](https://mintcdn.com/hamsa-452fa79a/etUfiFCo8p_6XpY_/images/media/voice-overs/video-config-light.webp?fit=max\&auto=format\&n=etUfiFCo8p_6XpY_\&q=85\&s=df0379ad3a228b8e250c4f7f726df8b8)

### Presets

| Preset | Settings |
| - | - |
| **Social reel** | Portrait 9:16, Cartoon, Word by word captions, Fade, Ken Burns on |
| **Explainer** | Landscape 16:9, Infographic, Build up captions, Fade, Ken Burns off |
| **Story** | Portrait 9:16, Realistic, Full sentence captions, Dissolve, Ken Burns on |

### Video settings

| Setting | What it does | First time |
| - | - | - |
| **Visual style** | **Realistic**, **Cartoon** or **Infographic**. Cartoon draws in one of a few illustrated looks, picked for each video. | Realistic |
| **Aspect ratio** | **Portrait (9:16)**, 1080 by 1920, for social. **Landscape (16:9)**, 1920 by 1080, for wide screens. | Portrait (9:16) |
| **Ken Burns effect** | "Slow zoom and pan on images for dynamic motion." | Off |
| **Captions** | "Add synchronized subtitles to your video", from your script. | On |
| **Caption style** | **Word by word** highlights the word being spoken. **Build up** adds words one at a time. **Full sentence** shows the whole sentence. Shown when captions are on. | Word by word |
| **Caption box** | A background box behind the captions, easier to read. Shown when captions are on. | On |
| **Scene transition** | How one scene flows into the next: **None**, **Fade**, **Dissolve**, **Slide left**, **Slide right**, **Circle open**, **Wipe left**, **Zoom in**. | Fade |
| **Reference image** (**Advanced**) | "Optional. Used as a visual anchor for the AI." | None |
| **Design instructions** (**Advanced**) | Color palette, mood, brand feel, anything to keep the same in every scene. | Empty |

After your first video, the dialog opens with the settings you used last, in this browser. The reference image and design instructions start empty each time.

### What AI makes

The video follows your take: one scene per sentence of the script, timed to the voice, with an AI image for each scene and your audio underneath. Images keep recurring people and places consistent from scene to scene, and contain no readable text, logos or watermarks; the only text on screen is the captions. Every video carries a small Hamsa logo in the top left corner.

## Good to know

* In **Prompt**, the three options are three different scripts, each with its voice and, with music on, its own track. When fewer than three voices fit, options can share a voice.
* **My script** has no dialect menu and no file: the voice sets the dialect, and the script is read word for word.
* Changing project while a file is uploading stops it, and no voice over is created.
* A video can only be made from a voice over whose take you kept.

## Related

> [card] [Manage voice overs](https://docs.tryhamsa.com/media/voice-overs/manage-voice-overs.md)
>
> History, downloads, videos and picture in picture.

> [card] [Voice overs](https://docs.tryhamsa.com/media/voice-overs/overview.md)
>
> Prompt vs My script, use cases and credits.

> [card] [Project defaults](https://docs.tryhamsa.com/media/workspace/project-defaults.md)
>
> Start every voice over in the same dialect.

> [card] [Add a voice](https://docs.tryhamsa.com/media/voices/add-a-voice.md)
>
> Clone a voice to read your scripts.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.