Skip to content

Product Overview

Facebetter is a real-time beauty effects engine designed for developers. It delivers high-quality AI-powered beauty processing for both images and live video streams. Built on a GPU rendering pipeline, the engine achieves low latency, minimal CPU usage, and high frame rates, and integrates seamlessly into Android, iOS, macOS, Web, and other major platforms.

Core Features

Skin Beauty

Fine-grained skin enhancement with four independently adjustable parameters — smoothing, whitening, rosy tone, and sharpening. Each parameter supports an intensity range of 0–100 for precise control over the final look.

Face Reshaping

Comprehensive adjustments spanning facial contour and feature proportions:

  • Face shape: face slimming, V-face, narrow face, short face
  • Contour: cheekbone slimming, jaw slimming, chin slimming
  • Features: nose bridge slimming, eye enlarging, eye distance adjustment

A single frame supports up to 5 people being reshaped simultaneously.

Makeup

Realistic layered makeup effects including lipstick (color and opacity adjustable) and blush (multiple preset shades). Multiple makeup items can be applied at the same time.

A single frame supports up to 5 people with makeup applied simultaneously.

Filters & Styles

20+ real-time style filters based on LUT (Look-Up Table) color grading, covering portrait, food, landscape, modern, and other themes. Intensity adjustment and custom LUT extension are supported.

Virtual Background

Powered by real-time portrait segmentation:

  • Background blur: keeps the foreground subject sharp while applying Gaussian blur to the background
  • Background replacement: swaps the background with any image asset

Sticker Effects

2D sticker effects including:

  • Screen stickers: overlaid at a fixed position on the frame
  • Face-tracking stickers: anchored to face keypoints and dynamically follow head movement

Sticker resources are distributed as .fbpack packages with support for hot-loading and dynamic switching.

Advanced Features

  • Face keypoint callbacks: returns 106 facial landmark coordinates per frame for custom effect development and business logic
  • Green screen keying (Chroma Key): precise removal of green backgrounds, suitable for live streaming and video compositing

Input / Output Formats

Image Formats

TypeSupported Formats
Static imagesJPEG, PNG, BMP
Video framesYUV I420, NV12, NV21, RGB, RGBA, BGR, BGRA
GPU textureOpenGL / OpenGL ES Texture

Resolution Limits

ScenarioMaximum
Real-time video stream720P @ 60fps / 1080P @ 30fps
Static image processingUp to 4K

Lossless conversion between all supported formats is available through a unified interface.


Performance

MetricValue
Render latency~10ms average at 1080P (GPU parallel)
CPU usage< 5% during real-time rendering
Memory footprintPrimarily GPU memory; system RAM usage is low

Supported Platforms

PlatformOS / Version
MobileiOS 10.0+, Android 5.0+, HarmonyOS 5.0+
DesktopmacOS 10.13+, Windows 10+, Linux
WebWebAssembly-based; compatible with Chrome, Firefox, Safari, and other modern browsers

Integration Languages & Interfaces

InterfaceUse Case
Java / KotlinAndroid applications
Objective-C / SwiftiOS and macOS applications
C / C++Cross-platform desktop (Windows, Linux, macOS)
JavaScript (WASM)Web applications (browser and Node.js)

Quick Start

Select your target platform and follow the integration guide: