AutoBunny

Overview

AutoBunny is a dealer-focused vehicle marketplace built as a heavily customised Laravel e-commerce platform. The problem it solves is specific to the auto trade: dealers list inventory in bulk, buyers hunt for very particular specifications, and a lot of stock moves through timed auctions rather than fixed-price sales. A generic storefront handles none of that well.

Key Features

• Timed auctions with automated settlement: Queued jobs resolve auctions on schedule, determine winners, and trigger the downstream notifications — no manual close-out.
• Dealer bulk upload: Dealers import inventory from spreadsheets rather than listing vehicles one at a time, with a background job pipeline handling the parsing and a separate queue processing product images so large imports don’t block the request cycle.
• Saved searches and alerts: Buyers save a search specification once and get notified when matching stock appears, via daily and weekly alert jobs.
• Push notifications: Firebase integration delivering auction and listing alerts to buyers on mobile.
• Multi-vendor architecture with dealer and seller roles: Separate flows for dealer listings versus private sellers, with role-appropriate dashboards and permissions.
• REST API layer: Dedicated API controllers for products, customers, sellers, search, and social login, backing the mobile experience alongside the web storefront.

Technologies Used

• Laravel: Application framework, queues, scheduled commands, and job pipeline.
• Bagisto: Open-source Laravel e-commerce foundation, extended well past its stock multi-vendor behaviour.
• Firebase Cloud Messaging: Push delivery for alerts and auction events.
• Queued jobs & scheduled commands: Auction settlement, winner updates, bulk listing imports, image processing, and recurring product alerts.
• Spreadsheet ingestion: Console commands for reading and normalising dealer inventory data.

Challenges and Learnings

Auctions are unforgiving about timing and idempotency. A settlement job that runs twice, or a notification that fires after a winner has already been recorded, produces disputes rather than bugs — so the queue work had to be safe to retry and correct under concurrency. The bulk upload pipeline taught a related lesson: splitting listing creation from image processing into separate queues kept a dealer importing 500 vehicles from degrading the site for everyone else.

Outcome

AutoBunny gave dealers a marketplace that matches how the auto trade actually operates — bulk inventory in, auctions and fixed-price listings out, with buyers kept in the loop automatically rather than having to keep checking back.