RecyclAIble
A computer-vision mechanical waste bin that autonomously separates trash and recycling, built as a multi-hackathon hardware/software project.
- Repo: https://github.com/Kevin-Liu-01/RecyclAIble
- Live site: https://recyclaible.vercel.app
- Devpost: https://devpost.com/software/recycleai-v1dqz5
- Awards: Best Hardware Hack at PennApps XXIII Sustainability; 1st Place at EcoHacks; additional Devpost-listed wins/prizes. Source: raw/articles/devpost-recyclaible-2026-06-25.md; raw/github/kevin-repos-2026-06-25.json
What it does
RecyclAIble detects when an item is placed on a bin lid, captures an image, classifies whether the item is recyclable, then tilts the lid into the appropriate trash or recycling compartment. The web interface shows recycling statistics and compares usage against broader averages. Source: raw/articles/devpost-recyclaible-2026-06-25.md
Implementation
The hardware used cardboard, a Raspberry Pi 3 B+, ultrasonic sensor, servo motor, and Logitech USB webcam. The backend accepted base64 images from the hardware through Flask, decoded them, sent them through a TensorFlow convolutional model, stored data in SQLite, and returned the classification to the Raspberry Pi. The frontend was a responsive Next.js/React/Tailwind interface. Source: raw/articles/devpost-recyclaible-2026-06-25.md
Kevin's listed role on Devpost: he developed the frontend that users interact with. The team roles separated AI model work, hardware/image capture, frontend, backend/database, and promotional materials. Source: raw/articles/devpost-recyclaible-2026-06-25.md
Why it matters
RecyclAIble is the clearest early example of Kevin shipping across hardware, AI, backend, and frontend in a hackathon setting. It also carries unusually strong public proof for an early project: Devpost artifact, visible hardware screenshots, awards, project story, implementation stack, and deployment. Source: raw/articles/devpost-recyclaible-2026-06-25.md
Current Evidence (2026-07-01)
This page was refreshed against Kevin's live GitHub inventory, pinned repositories, PortfolioMon cards, README scrape, and downloaded portfolio screenshots. Treat this section as the current evidence anchor; deeper interpretation stays in the surrounding compiled truth.
| PortfolioMon card | Types | Site | Screenshot |
|---|---|---|---|
| #031 RecyclAIble | AI/Hardware | site |  |
- PortfolioMon card #031; AI/Hardware; HP 290, ATK 135, DEF 90, SPD 80; standard card. Smart recycling system that won 1st Place in Hardware at PennApps XXIII. Uses OpenCV object detection on a Raspberry Pi to identify waste items in real time, then activates servo-controlled bins to sort recyclables, compost, and trash automatically.
| GitHub repo | Visibility | Language | Stars | Last push | Homepage |
|---|---|---|---|---|---|
| Kevin-Liu-01/RecyclAIble | public | TypeScript | 1 | 2026-06-16 | site |
Timeline
-
2026-07-01 | Project graph refresh checked this page against Kevin's live GitHub inventory, pinned repos, scraped PortfolioMon cards, READMEs, and portfolio screenshots; current evidence was added above. Source: raw/github/kevin-repos-2026-07-01.json; raw/github/kevin-pinned-repos-2026-07-01.json; raw/sites/kevinliu-biz-2026-07-01/portfolio-projects.json
-
2022-09-03 | GitHub repo created. Source: raw/github/kevin-repos-2026-06-25.json
-
2022-09 | Devpost project submitted; awards included PennApps XXIII Best Hardware Hack and EcoHacks 1st Place. Source: raw/articles/devpost-recyclaible-2026-06-25.md
-
2026-06-25 | Deployment verified at https://recyclaible.vercel.app and page created from Devpost/GitHub evidence. Source: raw/github/deployment-check-2026-06-25.tsv