hinge-liker
by mattttman
Image & Video GenerationBrief
Automated Hinge dating profile liker using Android emulator + Gemini vision AI.
Description
What This Skill Does
Automates Hinge swiping on an Android emulator. Uses Gemini vision AI to evaluate profiles, select the best photo or prompt, and send a generated witty comment. After each session, outputs a detailed report of profiles seen, liked, skipped, and comments sent.
Combines Android emulator automation with vision AI to make context-aware like decisions rather than random or threshold-based swiping.
When to Use It
- Running daily Hinge likes without manual swiping
- Letting AI pick the best photo to like on a profile
- Getting a session summary of who was liked and why
- Scheduling automated swipe sessions via cron
- Reviewing skipped profiles with AI-generated reasoning
Example Workflow
User asks: run 8 Hinge likes for me
- 1
Boots the Android emulator with the HingePhone AVD
- 2
Launches the Hinge app via adb shell am start
- 3
Runs hinge_android.py with --likes 8 and user description
- 4
Gemini vision evaluates each profile and selects the best photo or prompt to like
- 5
Sends witty AI-generated comment and logs result for each profile
Session report: 12 profiles seen, 8 liked, 4 skipped — includes who each liked person is, the comment sent, and skip reasons for the rest
Requirements
- GEMINI_API_KEY environment variable
- Android emulator (AVD) with Hinge installed and logged in
- Java runtime for Android emulator
- adb in PATH
Installation
clawhub install mattttman/hinge-likerRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.