muapi Logomuapi Logo
muapi Logomuapi Logo

Depth Anything V2: Editor de vídeo con IA

Extract depth maps con Depth Anything V2. Turn any imagen o vídeo en a per-frame structure signal fo AI vídeo motion transfer. Pruébalo gratis. Depth Anything V2 extracts a precise per-pixel depth map de a single imagen o every frame of a vídeo, producing a grayscale map where brightness encodes relative distance de the camera. Para vídeo entrada, frames are processed in batches on GPU y reassembled at the source resolution y frame rate, giving a temporally aligned depth sequence. This is commonly used as a motion y structure control signal fo AI vídeo generación models such as Seedance 2 y Seedance 2.5, where a driving vídeo's depth map guides camera movement y scene geometry mientras the appearance is generado fresh. It is equally useful fo 3D-aware compositing, relighting, background removal, y any pipeline that needs a stable geometric reference de ordinary RGB footage.

Controles interactivos del modelo

Extract depth maps con Depth Anything V2. Turn any imagen o vídeo en a per-frame structure signal fo AI vídeo motion transfer. Pruébalo gratis.

📝

Descripción general

Acerca de este modelo

Depth Anything V2 extracts a precise per-pixel depth map de a single imagen o every frame of a vídeo, producing a grayscale map where brightness encodes relative distance de the camera. Para vídeo entrada, frames are processed in batches on GPU y reassembled at the source resolution y frame rate, giving a temporally aligned depth sequence. This is commonly used as a motion y structure control signal fo AI vídeo generación models such as Seedance 2 y Seedance 2.5, where a driving vídeo's depth map guides camera movement y scene geometry mientras the appearance is generado fresh. It is equally useful fo 3D-aware compositing, relighting, background removal, y any pipeline that needs a stable geometric reference de ordinary RGB footage.

1Motion Transfer: Extract a depth vídeo de a driving clip to control camera y scene motion in Seedance 2/2.5 vídeo generación.
23D-Aware Compositing: Usa per-pixel depth to composite, relight, o apply depth-of-field effects on ordinary footage.
3Background y Foreground Separation: Isolate subjects de backgrounds using relative depth instead of green-screen segmentation.
4Robotics y Spatial Analysis: Estimate relative scene geometry de a single camera fo prototyping y research.
💰

Precios y tarifas

Análisis de costes

muapiapp$0.04/sec of video, or $0.05 flat per image

Pay por segundo of entrada vídeo processed, o a flat rate fo single-imagen depth maps, con sin suscripción.

Fal.aiNo disponible

Esta variante del modelo no aparece en Fal.ai.

ReplicateNo disponible

Esta variante del modelo no aparece en Replicate.

** Los precios de competidores son estimaciones basadas en arquitecturas de modelos y niveles de uso similares.

⚙️

Especificaciones técnicas

Esquema de configuración

URL de vídeostring

The URL of the driving video to extract a per-frame depth map from. Provide exactly one of video_url or image_url.

Valor por defectohttps://storage.googleapis.com/gtv-videos-bucket/sample/ForBiggerBlazes.mp4
URL de imagenstring

The URL of a single image to generate a depth map from. Provide exactly one of video_url or image_url.

Valor por defectohttps://d3adwkbyhxyrtq.cloudfront.net/webassets/videomodels/seedream-5.0-edit.jpg
📖

Guía de integración

Documentación para desarrolladores

Cómo usar Depth Anything V2 API

  1. Choose your entrada: Provide exactly one of video_url (a driving vídeo fo per-frame depth extraction) o image_url (a single frame fo a one-off depth map).
  2. Envía la solicitud: Send a POST request to /api/v1/depth-anything con a publicly accessible URL.
  3. Consulta el resultado o usa un webhook: The task runs asynchronously; poll the returned request ID o supply a webhook_url to be notified on completion.
  4. Feed the resultado downstream: Usa the returned depth map/vídeo as the motion o structure control entrada fo a vídeo generación model such as Seedance 2 o 2.5, o fo any depth-aware compositing pipeline.
curl -X POST "https://api.muapi.ai/api/v1/depth-anything" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "video_url": "https://example.com/driving-video.mp4"
  }'
  1. Recupera el resultado: Read the completed depth map URL (PNG fo imagen entrada, MP4 fo vídeo entrada) de the standard prediction result.
❓

Preguntas frecuentes

Preguntas frecuentes

¿Qué es Depth Anything V2?

It is a monocular depth estimation model that produces a grayscale depth map de a single imagen o every frame of a vídeo, where pixel brightness represents relative distance de the camera.

¿I provide vídeo_url o imagen_url?

Provide exactly one. Usa vídeo_url to extract a per-frame depth vídeo (e.g. fo motion-transfer control signals), o imagen_url fo a single-frame depth map.

What resultado format do I get back?

A vídeo entrada returns a grayscale depth-map vídeo at the same resolution y frame rate as the source. An imagen entrada returns a single grayscale depth-map PNG.

¿Qué es this used fo con Seedance 2 o 2.5?

Seedance's motion-transfer workflows accept a depth vídeo as a structure/control de movimiento signal. Extract the depth map de a driving vídeo con this API, then pass it en Seedance to guide camera movement y scene geometry mientras generating new appearance.

¿there a size o length limit?

La API acepta standard web-hosted imagen y vídeo files. Very long vídeos increase processing time since every frame is run through the depth model; keep driving vídeos to the length you actually need fo the downstream generación.

¿Cómo se calcula el precio?

Video entrada is billed at $0.04 por segundo of entrada vídeo. A single imagen entrada is billed at a flat $0.05 per request.