What it does

Auto-Tinder is a Python pipeline that automates the entire Tinder experience: reverse-engineering the web app’s internal API, scraping nearby profiles, classifying photos with your own manual labels, then retraining a deep CNN (Inception v3 via TensorFlow’s object detection API) to swipe on your behalf. It includes a full API wrapper with Person.like() and Person.dislike() methods, plus preprocessing that crops humans out of photos before training.

The interesting bit

The README doubles as a surprisingly thorough tutorial in API reconnaissance—complete with Chrome DevTools screenshots, Postman validation, and the discovery that Tinder stores original-resolution photos on publicly accessible URLs. The author also notes that gender remains API-visible even when hidden on profiles, which is either a useful dating insight or a privacy footnote, depending on your perspective.

Key highlights

• Reverse-engineers api.gotinder.com endpoints including /v2/recs/core, /like/{id}, and /pass/{id}
• Includes a complete Python API wrapper with requests and a Person class
• Uses TensorFlow Object Detection API to crop humans from profile photos before classification
• Retrains Inception v3 on user-labeled data for personalized swipe predictions
• Author explicitly warns this violates Tinder’s ToS and should not be used on real accounts

Caveats

• README is truncated mid-code-sample; the full implementation isn’t visible in the source provided
• Requires manual image labeling with a “simple mouse-click classifier”—the 5 hours of labeling the author jokes about is real unpaid labor
• API endpoints and auth mechanisms described are from an older web version; likely brittle against Tinder’s actual current defenses

Verdict

Worth a skim if you’re learning API reverse engineering or want a concrete, slightly absurd TensorFlow retraining example. Skip it if you’re looking for a maintained bot or have ethical qualms about automating dating platforms—even “for fun.”