agenthub · video.generate
Generate Video
Turn a text prompt into a video clip. Billed per second of video, not per token.
Why the node earns its place
The video counterpart to Generate Image, provider-agnostic in the same way, with the dropdown filtered to text-to-video models.
It is deliberately kept apart from Avatar Video, which is the talking-head path and requires a source image to animate. Keeping them separate means a model dropdown never offers a model whose required inputs the node cannot supply — the classic way a media node fails at the provider instead of on the canvas.
How it works at run time
- 1Take the prompt, and optionally an image to seed from.
- 2Call the text-to-video capability with the duration, aspect ratio and resolution tier.
- 3Return the clip URL, the full array, the duration and a usage descriptor, then emit a video-usage event for per-second billing.
What you wire
Inputs
- prompt *string
- imagestring
Outputs
- video_urlstring
- videosarray
- durationnumber
- usageobject
Configuration
| Key | Type | Default & options | What it does |
|---|---|---|---|
| provider * | string | choices from model-library:providers | Provider of the video model (e.g. google, openai, runway) |
| model * | string | choices from model-library:text_to_video_models | Text-to-video model |
| duration_seconds | number | default 8 | Requested clip length in seconds (billing is per second). Veo accepts only 4, 6, or 8; other values are snapped to the nearest. |
| aspect_ratio | string | default "16:9" 16:9 · 9:16 · 1:1 | Frame aspect ratio |
| resolution | string | default "720p" 720p · 1080p | Requested resolution tier (affects the per-second rate) |
| negative_prompt | string | — | What the model should avoid rendering |
What usually goes wrong
Watch for this
Duration and resolution both move the price, and duration is snapped to what the model accepts — some take only 4, 6 or 8 seconds, so a request for 5 becomes something else. Set negative_prompt for things to avoid rendering rather than trying to phrase the avoidance positively in the main prompt.
Behaviour & provenance