AI Systems
HiBid Auction Bot: Fully Autonomous Bidding Across 3 Live Auction Platforms — Zero Manual Intervention
Autonomous bidding across Rivercity, Sacramento, and Delaney HiBid auctions — real time, zero manual input. Proxy rotation, Max Speed surge mode, and live concurrency dashboard.
Django
PostgreSQL
Python
React
Redis
3
Auction Platforms Integrated (Rivercity, Sacramento, Delaney)
GraphQL
Full HiBid GraphQL LotBid API Integration
Max Speed
Burst Bidding Mode for Auction-Close Surge
The Problem
Live Auctions Close in Seconds — Manual Bidding Across 3 Platforms Is Impossible to Scale
Auto auction platforms like HiBid run hundreds of simultaneous lots across multiple locations, with soft-close timers that extend as bids come in. Manual bidding across three separate platforms — each with different authentication systems, different APIs, and different lot data formats — was impossible to scale. By the time a human operator placed bids across dozens of lots, the auction had already moved on.
The client needed a system that could watch every lot in real time, know the exact current price, and place bids instantly and autonomously — across all three platforms simultaneously — without any human involvement.
Our Solution
GraphQL + REST Bidding with Playwright Auth, Proxy Rotation, and Real-Time WebSocket Dashboard
We built HiBid Auto Auction Bidding Bot as a full-stack Django application with a React dashboard. The authentication layer uses Playwright headless Chromium to navigate each platform's login flow, capture session tokens (Bearer JWT for Rivercity via network interception, AWS WAF tokens and session cookies for Sacramento/Delaney), and store them in the Account table for all subsequent API calls.
The lot scrapers paginate through each platform's listing API efficiently scanning hundreds of live lots simultaneously, matching against imported watchlists and bulk-updating all matched records instantly. The bidding engine runs continuously: it queries all lots where current_amount < max_bid, randomly selects up to run_proccess lots, randomly assigns accounts, and dispatches all bids in parallel.
Rivercity bids use the full LotBid GraphQL mutation, returning bid status, suggested amounts, soft close timers, and complete lot state. Sacramento and Delaney bids use REST POST with the captured session credentials. All bid requests rotate proxies (random port from the configured pool) and fake user agents.
Max Speed mode runs the bidding engine at peak power — enabling surge bidding in the final seconds of any auction close. Django Channels WebSockets broadcast live stats to the dashboard after every batch, with operators able to adjust run_proccess inline without leaving the dashboard.
Architecture
Django 4.2 + DRF API. Django Channels WebSockets for live dashboard updates. Celery + Redis for task queue. Playwright headless Chromium for session authentication and cookie harvesting. Concurrent lot scraping and bid dispatch via async execution. HiBid GraphQL API (LotBid mutation, LotSearch query). Sacramento/Delaney REST APIs. 2captcha for CAPTCHA bypass. Rotating proxy pool (host/port range managed via dashboard). PostgreSQL with bulk ORM operations. React + Reactstrap + ApexCharts dashboard.
The Result
Fully Autonomous Auction Coverage — Every Matching Lot Bid Instantly, Around the Clock
HiBid Auto Auction Bidding Bot replaced an entire manual monitoring and bidding operation with a fully autonomous system that requires zero human involvement. Lots are tracked in real time across all three HiBid platforms simultaneously — the moment a current price drops below a configured maximum bid threshold, the bot places the bid instantly.
Max Speed mode ensured the system could handle the highest-competition auction closes — placing bids at full concurrency as close timers approached, without any operator action. Proxy rotation and user-agent rotation maintained uninterrupted platform access throughout. The real-time WebSocket dashboard gave operators complete visibility: lot counts, completion rates, live bidding history, and inline concurrency control — all without leaving the browser.
The client went from manual monitoring of individual lots to automated, around-the-clock coverage of every matching lot across three platforms simultaneously — with no additional headcount required.
Want scalable architecture and measurable results like this?
Book an Architecture Audit
Build Something Like This
// free scoping call · senior engineers · no commitment
Project scope, timelines, and architecture options in one 45-minute session.
More Work