Personalize Videos at Scale with a CSV

Before you begin

A Personalization project is intended for a video with one visible face.

Important: Use a Multi-speaker project whenever more than one face appears at any point in the source video, even when only one person speaks.

For best results:

• Use a clear video with the speaker facing the camera.
• Keep the personalized portions concise and clearly spoken.
• Review the video and CSV requirements in Prepare Your Content.
• Confirm that you have permission to use the speaker’s voice and likeness.

The Personalization workflow recommends a concise source video of approximately one to two minutes, although this is not a strict platform-wide duration limit.

How projects and campaigns are organized

A Personalization project can contain multiple campaigns. Each campaign has its own:

• Source video
• Recipient CSV
• Personalized variables
• Generated results

A Personalization project can currently contain up to five campaigns. There is no limit on the number of recipient rows within an individual campaign.

Create a Personalization project

1. From Home, select Create Personalization.
2. 1. You can also open Projects, select Create Project, and choose Personalization project.
2. On the All Campaigns page, select New Campaign.
3. Enter a descriptive campaign name.
4. Select Create campaign.

Each campaign follows four steps:

1. Upload Video
2. Add Recipients
3. Highlight Variables
4. Generate Results

Step 1: Upload the source video

Upload the video you want to personalize.

LipDub will process and transcribe the video. When Transcription completed appears:

1. Preview the uploaded video to confirm that it is the correct file.
2. Confirm that the speaker’s dialogue is clear.
3. Select Next.

The surrounding source audio— including music, ambience, and sound effects— is preserved. Only the words or phrases assigned as variables are replaced.

Step 2: Add recipients with a CSV

Upload the CSV containing the personalized data for each version.

Each row represents one generated video. Each column represents a field that can be assigned to a word or phrase in the source transcript.

Your CSV must:

• Use simple, unformatted column headers.
• Include a required email column.
• Contain a value in every cell.
• Have no blank rows or blank values.
• Contain one recipient or output version per row.

The email values do not need to be valid or unique, but every row must contain one. LipDub does not contact these addresses or send videos automatically. The email column is used to organize results and make exported video links easier to import into an external email platform.

There is no campaign row limit. A CSV can contain thousands of recipients and multiple data columns.

After the CSV is uploaded:

1. Confirm that CSV loaded successfully appears.
2. Review the Data Preview. LipDub displays the first five rows of the file.
3. Check that the column headers and values appear in the correct fields.
4. Select Next.

Tip: To control the pronunciation of a name or other value, enter a phonetic spelling in the CSV. Review the generated result before distributing it.

Step 3: Review the transcript

Before creating variables, read the complete Video transcript and correct any transcription errors.

Check especially:

• The words or phrases you intend to personalize
• Names
• Company and product names
• Numbers
• Abbreviations
• Punctuation around each personalized phrase

Use the edit control in the top right of the transcript to make corrections.

Important: Transcript editing is disabled after you create the first variable. Complete all transcript corrections before highlighting and assigning any variables. Additionally, note that editing your text will not modify the source audio.

Highlight and assign variables

To create a variable:

1. Highlight the exact word or phrase you want to replace in the transcript.
2. In the Create variable window, open Select CSV column.
3. Choose the CSV column that should replace the highlighted text.
4. Confirm that the variable appears under Selected Variables.

For example, you might highlight the name spoken in the source video and assign it to a CSV column named Name. LipDub will use the corresponding value from each row when creating that recipient’s video.

You can create up to 10 variable selections in one campaign.

The same CSV column can be assigned to more than one selection. For example, when the recipient’s name is spoken twice in the video, assign the Name column to both occurrences.

Under Selected Variables, you can:

• Review each variable and its source text
• Change its CSV-column assignment
• Delete a variable

Repeat the process until every word or phrase you want to personalize has been assigned.

Before selecting Next, confirm that:

• Every intended phrase has been highlighted.
• Each phrase is connected to the correct CSV column.
• The number of selected variables does not exceed 10.
• Every row in the CSV contains a value for the assigned columns.

Step 4: Generate the personalized videos

The Generate Results page lists the recipients and variable data loaded from the CSV.

Before starting:

1. Confirm the number of Recipients.
2. Confirm the number of Variables.
3. Review the credit estimate shown beside Generate results.
4. Select Generate results.

LipDub creates one personalized video for every row in the CSV. The page displays the status of each result while generation is in progress.

You can leave the page and return to the campaign later.

Preview the results

When generation is complete:

1. Select a recipient row.
2. Play the video in the preview panel.
3. Listen closely to each personalized word or phrase.
4. Check the pronunciation, timing, audio transition, and lip-sync.
5. Review several different rows before distributing the campaign.

Use Search videos to find a specific recipient in a larger campaign.

The results summary shows:

• Total recipients
• Number of variables
• Number of successfully generated results

When a value is not pronounced as intended, revise its spelling phonetically in the source CSV and update the campaign before generating the results again.

Tip: Instead of regenerating the whole batch, try creating a new campaign using a smaller CSV of just the affected pronunciations.

Download or share a video

Each completed recipient row provides two delivery options.

Copy the video link

Select Copy link to video to copy the recipient’s direct viewing link.

Personalization links are:

• Publicly viewable by anyone who has the link
• Unlisted
• Not indexed or publicly listed by LipDub
• Active while your LipDub account and the associated project remain active

Deleting the Personalization project also deletes its source files and outputs. Any links belonging to that project will stop working.

Custom URLs are not currently supported.

Download the video

Select the download control beside a recipient to save that video directly.

Downloaded videos can be distributed through another platform, including:

• Email
• Social media
• A CRM
• A sales-engagement platform
• Your own file-hosting or delivery system

Custom output filenames are not currently supported.

Export all video links as a CSV

Select Export CSV to download a spreadsheet containing the campaign’s recipient data and generated video links.

You can import this file into an external email, CRM, or campaign tool and match each video link to the corresponding email value.

Note: LipDub does not send the emails. Exporting the CSV gives you the data needed to distribute the videos through your preferred platform.

Check the campaign before distribution

Before sending or publishing the results, confirm that:

• Every expected CSV row produced a video.
• Each recipient received the correct variable values.
• Names and specialized terms are pronounced correctly.
• The personalized speech transitions naturally into the surrounding source audio.
• The lip-sync looks natural for each replacement.
• Every copied or exported link opens the correct video.

If one processing, training, or generation action remains in progress for longer than expected, contact support@lipdub.ai