Kris Garg Portfolio
100% On-Device • No Internet Permission

Sticker Search

Find the exact WhatsApp sticker you want in seconds — by the words on it, what it depicts, or who is in it — and drop it directly into your chat with a single tap.

Package: app.stickerfindAndroid 10+ (API 29–36)Kotlin • Jetpack • ONNX

Search Modes & Superpowers

Powered by lightweight on-device vision and language models

Multilingual Text Search

ML Kit Latin & Devanagari OCR scans every sticker text locally. Search memes in English, Hindi, or Hinglish with instant retrieval.

Semantic Picture Search

Powered by MobileCLIP-S0 running on ONNX Runtime. Search natural descriptions like “crying cat”, “happy birthday”, or “tea party”.

Face Recognition & People

ML Kit Face Detection + OpenCV SFace clusters faces of friends and family. Tag or name someone once, and find all stickers with their face.

Native WhatsApp Experience

Send stickers with one tap directly into chat

Android only permits keyboards to inject images into chat boxes, and accessibility helpers to switch keyboards. Sticker Search pairs a transparent helper with an input method so you never have to copy-paste or switch apps manually.

  • Round bubble button: Sits over your keyboard in WhatsApp, opens instant sticker search.
  • Text selection menu: Type a keyword, highlight it, and tap “Find sticker”.
  • WhatsApp Rule Fitting: Auto-formats animated stickers to 512×512 and <500KB so they never freeze.
Technical Architecture
LanguageKotlin 1.9 • Android SDK 36
OCR EngineGoogle ML Kit (Latin + Devanagari)
Vision ModelMobileCLIP-S0 (ONNX Runtime 1.20)
Face RecognitionML Kit Face + OpenCV SFace
Animation EngineCustom WebP Frame Transcoder
Network PermissionsNone (android.permission.INTERNET removed)

Built with Privacy by Design

Sticker Search has zero internet access. No server, no telemetry, no tracking. Read our complete privacy declaration filed for Google Play Store compliance.