"video_*" in the agent’s allowedTools array.
video_generate — Generate Video
Generate a video from a text prompt and save it as MP4.
Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
prompt | string | yes | Text prompt describing the video to generate |
path | string | yes | Output file path (e.g. output.mp4) |
model | string | no | fal.ai video model ID (default: fal-ai/wan/v2.2-1.3b/text-to-video) |
num_frames | number | no | Number of frames to generate (default: 81, ~5 seconds at 16fps) |
resolution | string | no | Video resolution as WIDTHxHEIGHT (default: 854x480) |
num_inference_steps | number | no | Inference steps — higher is better quality but slower (default: 30) |
guidance_scale | number | no | How closely to follow the prompt (default: 5.0) |
seed | number | no | Random seed for reproducible results |
Models
| Model | Quality | Speed |
|---|---|---|
fal-ai/wan/v2.2-1.3b/text-to-video | Good (default) | Faster (1-2 min) |
fal-ai/wan/v2.2-a14b/text-to-video | Best | Slower (3-5 min) |
Returns
Confirmation with file path, size, model used, and generation parameters.Notes
- Requires
FAL_KEYenvironment variable or a vault entry (service"fal", key"key") - Output is saved as MP4 within the agent’s sandbox
- Generation time varies from 1-5 minutes depending on model and resolution