MetaSAM 3 VideoVideo → Video

Segment and track objects through a video from a text prompt — name several objects separated by commas and SAM 3 keeps each one masked across every frame.

SegmentationObject TrackingVideo Input

Playground SDK API Advanced Usage Pricing

Install the SDK

The SDK targets Node 20+ and ships with TypeScript types.

npm install @picsart/ai-sdk

Authenticate

Create a client once with your API key — it's sent as a bearer token on every request.

import { createClient } from '@picsart/ai-sdk';

const ai = createClient({
  apiKey: process.env.PICSART_API_KEY, // sent as: Authorization: Bearer <key>
  apiUrl: 'https://api.picsart.com',
});

Run the model

Call the model with its parameters. This example uses the required ones.

import { createClient } from '@picsart/ai-sdk';

const ai = createClient({
  apiKey: process.env.PICSART_API_KEY,
  apiUrl: 'https://api.picsart.com',
});

const result = await ai.generate('sam-3-video', {
  prompt: "A serene mountain lake at golden hour, ultra detailed",
  videoUrl: "https://cdn.example.com/input.mp4"
});

console.log(result.url); // video URL

Parameters

5 parameters. Required ones must be supplied; the rest fall back to their defaults. The API takes these same names over HTTP.

Parameter	Type	Required	Default	Details
`prompt` Prompt	`text`	yes	`—`	e.g. "person, cloth"
`videoUrl` Source Video	`file (video)`	yes	`—`	—
`detectionThreshold` Detection Threshold	`range`	no	`0.5`	0–1 (step 0.05)
`applyMask` Apply Mask	`boolean`	no	`true`	—
`outputFormat` Format	`enum`	no	`mp4`	mp4, webm

Output

generate() resolves once the job completes. A video URL.

Result

{
  "items": [
    {
      "url": "https://cdn.picsart.com/…/result"
    }
  ],
  "model": "sam-3-video"
}