p-video
Pruna Video es un generador de vídeo rápido y multimodal — texto-a-vídeo, imagen-a-vídeo y audio-a-vídeo desde un único endpoint. Un **modo borrador** integrado reduce el coste a aproximadamente una cuarta parte para iterar rápido; desactívalo cuando estés listo para la toma final.
Entrada
Describe the video to generate. When `image` or `audio` is provided, the prompt directs how the scene develops from those inputs.
Optional first-frame image for image-to-video mode. Supports jpg, jpeg, png, webp.
Optional last-frame image. When paired with `image`, the model interpolates between the two key frames over the duration.
Optional audio track that conditions the generated video (flac / mp3 / wav). When provided, `duration` is ignored and the output length follows the audio.
Output length in seconds (1-20). Ignored when `audio` is provided.
Output aspect ratio. Ignored when an input `image` is provided — the image's aspect ratio is used.
Output video resolution. 1080p costs roughly 2× more per second than 720p.
Frames per second. 24 is the cinematic default; 48 is smoother for fast motion at the same duration.
Draft mode — generates a lower-quality preview for fast iteration at about a quarter of the standard cost.
When true, the output video keeps an audio track (either the input `audio` or model-generated audio). Set false for silent output.
Auto-expand the prompt for better motion direction. Recommended on; turn off for strict adherence to your wording.
When true, prompts and the input image skip safety screening before generation. On by default for this model.
Optional random seed for reproducible generations.
Salida
Ejecuta el modelo para ver el resultado aquí.
Precios
Características
Ejemplos de API
curl -X POST \
/api/v1/model/prunaai/p-video \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"input": {}
}'