2026-09-28

Faceless UGC: what it is and how to make it with AI

Faceless UGC is creator-style content showing hands, the product and text, with a voiceover instead of a face. How to make it with AI in three steps.

5 min readSume

All posts

Faceless UGC is creator-style content in which the camera shows hands, the product, a phone screen or text on screen, and a voiceover does the talking instead of a face on camera. To make it with AI, generate a hands-on product clip, generate the voiceover from a script, and put the two together in one edit.

The Sume steps below come from the Format catalog, Timeline 1.0 and the API reference for text to speech, read on 2026-09-28. Prices come from Sume's rate card.

What does faceless UGC look like?

The same shapes as creator UGC, minus the talking head. Common treatments:

A hands-on demo: the product being applied, poured, opened, or switched on, with the voiceover explaining it.
A point-of-view unboxing: the box, the opening, the reveal.
Close-ups of the product with the key claims as text on screen.
For apps, the phone screen in a hand while the voice walks through one task.

How is it different from AI UGC with a presenter?

In AI UGC with a presenter, a generated person speaks your script to camera; what is AI UGC covers that version. Faceless UGC drops the presenter, so the picture and the voice are made separately and joined afterwards. It is also different from faceless channels in the automation sense, which automating faceless short-form videos covers.

How do I make faceless UGC with AI?

Three calls, in this order. The edit takes only Sume-hosted media from this workspace, which the first two steps produce.

From the Format catalog, the API reference, and Timeline 1.0, read 2026-09-28.
Step	Call	What to send
1. Hands-on clips	`POST /v1/formats/sume/sume-product-usage-demo/runs`	Product photos in `attachments`; the action and "no face on camera" in `instruction`
2. Voiceover	`POST /v1/tts-1.0/generate`	The script as `transcript` (up to 20,000 characters) and a voice: `avatar_handle` or `voice.id`
3. The edit	`POST /v1/timeline-1.0/render`	The voiceover as the `audio` spine and the clips as ordered `video[]` slots

How do I put the clips and the voiceover together?

The product usage demo Format is described for "hands-on product ads", and its run output is a durable media.sume.com URL. In current code the voiceover's result carries its file as audio_url, a Sume-hosted artifact. Both can go straight into one Timeline 1.0 render, with the voiceover as the audio spine and the clips as video[] slots; the render call itself, with a music bed, is step 5 of automating faceless short-form videos.

One thing to plan for in a faceless ad: in current code a Timeline render keeps only the spine's sound (plus an optional soundtrack) and drops each clip's own audio. A pour or a click on the demo clip is gone in the edit, so let the voiceover and any soundtrack carry all the sound.

How much does faceless UGC cost to make?

Text to speech is $0.0475 per 1,000 characters and a Timeline render $0.10 per output minute, rounded up to whole minutes, both plus a 5.5% agent fee by default. The clips' cost depends on the models the Format's run uses; its generation_spend_cap_usd caps it, and the rest of the rate card is on API pricing.

What are the limits?

A Timeline render is 1–1,800 seconds long with 1–200 video slots, and every URL must already be this workspace's media.sume.com file, such as an earlier Sume job's output.
Set language on the voiceover for any non-English script; omitted, it defaults to English.
Captions: in current code the caption job refuses a source with no audio stream or one over 60 seconds. The finished edit carries the voiceover, so it has audio; split anything longer first.
To use your own voice, see AI voiceover in your own voice; step-by-step voiceover help is in how to make an AI voiceover.