Faceless Video Generator
Turns a topic prompt into a narrated vertical short with captions synced to the voice and an opening hook title. Six video models are available to render the scenes.
Overview
Faceless Video Generator is a Wiro app for YouTube Shorts, Instagram Reels, TikTok and X. From a topic, it writes a short script, reads it aloud in the voice you pick, renders the scenes with one of six video models and burns in captions that follow the narration. Each video opens on a large hook title. There is no presenter unless you add a talking head, though the generated scenes often show people.
What you can build
- Myth-versus-fact explainers, like the black hole and fitness samples
- Place comparisons such as the "Tokyo vs New York" sample
- Numbered fact lists or pros-and-cons pieces for a recurring series
Inputs
- A prompt with the topic and the angle.
- Video model: Wan2.2 (default), MiniMax Hailuo 2.3 Fast, MiniMax Hailuo 2.3, Google Veo 3.1 Fast, Google Veo 3.1 or ByteDance Seedance Pro v1.5.
- Length from 10 to 30 seconds in 5-second steps (default 10) and a quality setting of 480p, 720p or 1080p (default 480p), which renders 304x544, 408x720 or 720x1280 frames.
- 54 narrator voices (Liam by default) and speech speed from 0.5 to 2.0 (default 1.3).
- Caption effect (Typewriter, Sentence or the default Flash Each Word), position, Poppins weight, colors and box opacity.
- Hook title font and colors, with an optional zoom or pan.
- An optional talking head from an image URL.
Outputs
An MP4 with the voice and a music bed mixed together and the captions burned in. The five samples are 720x1280 (9:16), run 28 to 33 seconds and change scene every few seconds. Their captions build up in white and flash yellow on some lines.
Recommended settings
- Write the prompt as an angle ("six surprising facts about...", "X vs Y from a local's perspective"), as the example prompts do.
- Try topics and voices on Wan2.2 at 480p before picking the model and resolution for the final cut.
Limitations
- Check the script's facts. One sample says Istanbul's Beyazit Square holds the world's oldest public library.
- Scenes can contain garbled text, such as a banner misspelling "commuting" in the remote-work sample.
- Length follows the narration: three samples run past 30 seconds, the longest option.
Safety & compliance
If you add a talking head made from a photo of a real person, get that person's consent first.
Example prompts
Great starting points for Faceless-Video-Generator.
API quick start
Run Faceless-Video-Generator with a single API call.
{
"prompt": "Create a video comparing myths vs facts a…",
"videoModel": "Wan2.2",
"videoQuality": "480p",
"ttsVoice": "am_liam"
}curl -X POST "https://api.wiro.ai/v1/Run/wiro/faceless-video-generator" \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_WIRO_API_KEY" \
--data-binary @- <<'JSON'
{
"prompt": "Create a video comparing myths vs facts a…",
"videoModel": "Wan2.2",
"videoQuality": "480p",
"ttsVoice": "am_liam"
}
JSONOther models you might like
Explore similar video generation models.



