Features and limits
One API, two workflows, one output format. Everything below is what the pilot does today.
Two workflows
Script to video
script-to-videoYou send the exact text. It is narrated verbatim: no rewording, no added hook or call to action. Only leading and trailing whitespace is trimmed.
Prompt to video
prompt-to-videoYou send a topic and a target duration (15 to 60 seconds, 30 by default). A script is written for you, split into scenes, then narrated. The duration is a target, not a guarantee.
What goes into each video
Narration
A single English voice reads the script. Word timings from the voice provider drive the captions and the scene cuts.
Stock footage
Each scene gets a vertical clip searched on Pexels from its keywords, cropped to the frame. When no suitable clip is found, a plain colored background keeps the video complete.
Captions
Large burned-in captions, grouped in short lines and highlighted word by word, kept clear of the interface of vertical platforms. The text is the script, word for word.
Music
A background track mixed under the voice, enabled by default and disabled with music.enabled = false.
Format
Vertical 1080x1920 MP4, H.264 video and AAC audio at 30 frames per second. The layout is fixed by the service.
Retention
The MP4 stays downloadable for 7 days after it is ready, then it is deleted. Download links are short-lived and requested again through the API.
Public limits
| Item | Limit |
|---|---|
| Script or topic | 1 to 1000 characters after trimming |
| Target duration of a topic | 15 to 60 seconds |
| Client references (metadata) | Up to 20 string pairs, returned as is |
| Output | 1080x1920, 30 fps, MP4 |
| Availability of the file | 7 days after the video is ready |
| Language and voice | English, one voice |
Not in the pilot
- No dashboard or web form: videos are created through the API.
- No voice, theme or layout choice, and no editor.
- No webhooks: poll the video status instead.
- The MCP server for AI assistants is coming later.
Ready to integrate? The documentation covers authentication, creation, polling and download.