Media
Video Regional Effects
Apply visual effects to specific video regions with tracking: blur, grayscale, sepia, highlight, dim, transparency, etc. Works with VideoObjectDetection results. Inverse mode applies effects everywhere except regions. Note: Transparency effect requires WebP output or preserves MP4 format.
Category: Media · Plans: Trial, Pro, Premium, Enterprise · Type: VideoRegionalEffects
Inputs#
| Field | Type | Required | Default | What it does |
|---|---|---|---|---|
| Video URL | Text | Yes | (empty) | URL to the video to apply regional effects to. |
| Tracking Results (JSON) | Text | No | (empty) | JSON from VideoObjectDetection with tracked objects. Objects with “appearances” will be tracked. |
| Effect | Text | No | blur | Visual effect to apply to regions. Options: blur, grayscale, sepia, negative, highlight, dim, saturate, desaturate, solid_color, transparency |
| Blur Method (if effect=blur) | Text | No | gaussian | Method for blur effect. Options: gaussian, pixelate, solid (Advanced) |
| Effect Strength | Number | No | 0.5 | Strength/intensity of effect (meaning varies by effect type). |
| Inverse Mode | Yes/No | No | Off | If enabled, applies effect everywhere EXCEPT the tracked regions. |
| Expand Regions (px) | Number | No | 10 | Expand region boundaries by pixels for better coverage. (Advanced) |
Outputs#
| Output | Type |
|---|---|
| Video Url | Text |
| Duration Sec | Number |
| Frames Processed | Number |
| Regions Per Frame | Number |
| Effect Applied | Text |
| Inverse Mode | Yes/No |
| Status | Text |
| Error | Any |
Use as an agent tool#
Tool name: videoregionaleffects
Apply visual effects to tracked regions in videos (pairs with VideoObjectDetection). Use cases: blur faces for privacy, highlight detected objects, background effects, censor sensitive content. Supports 10 effects: blur (gaussian/pixelate/solid), grayscale, sepia, negative, highlight, dim, saturate, desaturate, solid_color, transparency. Inverse mode lets you affect everything except detected objects (e.g., blur background while keeping person sharp). Requires tracking_results JSON from VideoObjectDetection step.