SandBase is live — $1 in free credits on signupStart free ›
Start building
Use in agentMeta models

Meta modelsimage generation api

meta/sam-3/video-rle

Sam 3 Video Rle is Meta's video-to-video AI model. Transform, enhance, and edit video content using text prompts - from style changes to object manipulation and scene modification.

Input
Text prompt for segmentation. Use commas to track multiple objects (e.g., 'person, cloth').

PNG, JPEG, WebP, or GIF · 20 MiB maximum

The URL of the mask to be applied initially.
The URL of the video to be segmented.
Apply the mask on the video.
Return per-frame bounding box overlays as a zip archive.
List of box prompts with optional frame_index.
List of point prompts with frame indices.
0.011
Detection confidence threshold (0.0-1.0). Lower = more detections but less precise. Defaults: 0.5 for existing, 0.7 for new objects. Try 0.2-0.3 if text prompts fail. Range: 0.01 to 1.
Frame index used for initial interaction when mask_url is provided.
Idle

Example output — click Run to generate your own

API README

Sam 3 Video Rle

Sam 3 Video Rle is built for promptable segmentation and reconstruction, with this exact route focused on run-length encoded video tracking. Sam 3 Video Rle is Meta's video-to-video AI model. Transform, enhance, and edit video content using text prompts - from style changes to object manipulation and scene modification. It is designed for production workflows that need a clearly scoped media operation instead of a generic assistant response, preserving the supplied direction while returning an artifact ready for review, iteration, or downstream assembly.

The request surface keeps creative direction and delivery choices explicit through prompt, mask_url, sync_mode, video_url, apply_mask, box_prompts. Teams can lock source material, describe the intended result, choose supported output controls, and rerun the same request structure across variations. This makes experiments easier to compare and larger pipelines easier to operate without hiding important decisions inside an opaque preset.

Highlights

  • Run-Length Encoded Video Tracking. The meta/sam-3/video-rle route is dedicated to run-length encoded video tracking, keeping inputs and outputs matched to this operation.
  • Precise Spatial Selection. Within meta/sam-3/video-rle, documented prompt and media controls direct precise spatial selection while keeping the brief close to the result.
  • Structured Masks Or Geometry. The meta/sam-3/video-rle result provides structured masks or geometry in a machine-consumable form for review and downstream production.
  • Route-specific control. For meta/sam-3/video-rle, documented controls include prompt, mask_url, sync_mode, video_url, apply_mask, box_prompts, defining the supported variation space for this exact route.

Pricing

ConfigurationPrice
Per request$0.005000

When to Use

ScenarioWhy it fits
Choose this routeWhen the deliverable is specifically run-length encoded video tracking, not a neighboring generation or editing task.
Production iterationWhen creative teams need repeatable precise spatial selection with an explicit request payload.
Automated pipelinesWhen structured masks or geometry must flow into review, storage, compositing, or publishing.
Controlled variationsWhen a stable brief should be rerun while changing only documented inputs.
Prototype to scaleWhen the same meta/sam-3/video-rle contract should serve a single experiment and a larger batch later.

Prompt Guide

Lead with the subject or source, then state the transformation, important spatial or temporal details, style, and finish. Supply only documented fields, keep media URLs reachable, and vary one control at a time when comparing outputs.

{
  "prompt": ""
}

Technical Specs

PropertyValue
Model IDmeta/sam-3/video-rle
Required inputsprompt
promptstring; default:
mask_urlstring
sync_modeboolean; default: false
video_urlstring
apply_maskboolean; default: false
box_promptsarray; default: []
frame_indexinteger; default: 0
point_promptsarray; default: []
boundingbox_zipboolean; default: false
safety_tolerancestring; default: 6
detection_thresholdnumber; default: 0.5

Related Models

  • meta/sam-3/3d-align
  • meta/sam-3/3d-body
  • meta/sam-3/3d-objects
  • meta/sam-3/image

Related Models

meta/sam-3/3d-alignSam 3 3d Align by Meta - advanced AI model for 3d-to-3d. Delivers high-quality results with fast inference, suitable for both creative and production workflows.meta/sam-3/3d-bodySam 3 3d Body is Meta's image-to-3D AI model. Transform photographs into production-ready 3D meshes with accurate geometry and texture mapping.meta/sam-3/3d-objectsSam 3 3d Objects by Meta - convert 2D images into 3D models with AI. Generate textured 3D assets from single photos for games, AR/VR, e-commerce, and digital content creation.meta/sam-3/imageSam 3 Image is Meta's intelligent image editing model. Transform, retouch, and reimagine existing images using text prompts - from background replacement to artistic style conversion.meta/sam-3/image-rleSam 3 Image Rle is Meta's intelligent image editing model. Transform, retouch, and reimagine existing images using text prompts - from background replacement to artistic style conversion.meta/sam-3/image/embedSam 3 Image Embed by Meta - advanced AI model for vision. Delivers high-quality results with fast inference, suitable for both creative and production workflows.meta/sam-3-1/image-rleSam 3 1 Image Rle is Meta's intelligent image editing model. Transform, retouch, and reimagine existing images using text prompts - from background replacement to artistic style conversion.meta/sam-3-1/videoSam 3 1 Video is Meta's video-to-video AI model. Transform, enhance, and edit video content using text prompts - from style changes to object manipulation and scene modification.