How to Record a Reusable Video for Personalization

1. Plan the reusable message

Write a short script with a useful message for every recipient. Keep the greeting and main point fixed, and decide which details—such as a person’s name, company, or call to action—should change. BHuman uses a source video as the template, then replaces marked variables with values from campaign data.

• Keep changeable spoken phrases short, such as a name or company.
• Use distinct variable names for different kinds of information, such as first_name and company.
• Leave fixed words out of the recipient data. If “Hi” is already in the template, the first_name value should contain only the name.

2. Record the base video in BHuman

In AI Studio, name the template, choose Record in BHuman, allow camera and microphone access, and record your base take. Keep script notes beside the recorder if useful. Speak the variable phrases clearly, with a pause around each one.

• Use soft, even lighting and a stable camera position.
• Sit several feet from the camera so your face is not too close to the frame.
• Keep your head still during variable phrases.

3. Or upload an existing recording

If you already have an approved take, choose the upload route instead of recording in the browser. The uploaded video becomes the template source; you then continue through transcript, variable, voice, and preview setup. The browser recorder does not open in upload mode.

4. Mark the parts that will change

After recording or uploading, continue to the transcript and samples. Select the words or media slots that should vary by recipient, then give each variable a clear, stable name. Use lowercase snake_case, avoid duplicate names, and keep text and media values in separate variables.

• Example: in “Hi, {first_name}, welcome to {company},” the row values should be the recipient’s name and company—not the whole greeting.
• For spoken text, send only the words that replace the marked phrase; the fixed words remain in the template.
• For a variable background or call-to-action link, use a name that explains its purpose, such as asset_url or booking_url.

5. Preview with real sample values

Generate a few preview samples and check pronunciation, timing, subtitles, and any background assets before preparing a larger batch. Short variable phrases tend to have more predictable timing. Review transcript timing when subtitles or manual timing are used.

• Check that names and company phrases sound natural in the sentence.
• Confirm that each variable is mapped to the right source field and that no fixed words are repeated.
• If a face is too close for sample generation, AI Studio may ask you to choose a safer preview mode.

6. Prepare campaign data and generate

Start a campaign from a ready template with a source video and at least one variable. Upload a CSV or Excel file, connect a supported workflow, or use an API; then map each template variable to the matching column or payload field. Check that every required variable has a mapping and every row has a value before generating.

• For a CSV or Excel batch, use readable headers that match your variable names where possible.
• Preview the rows and correct blank or mismatched values before you submit the campaign.
• Generation runs asynchronously. Check the generated videos page for queued, processing, succeeded, failed, or needs-review results.

Choose recording or upload based on your starting point

Record in BHuman when you want to create the source take with its browser camera and microphone. Upload when you already have a suitable recording or an approved production take. Both routes lead into transcript, variable, and sample setup; the choice is about where your source video comes from.