Work Management System

When you have many workers doing the same kind of small task - like solving captchas by hand while the rest is automated - the hard part is not the task. It is running the team. Who does what? Who did how much? Who gets paid what? So I built the system that runs all of it.

Each worker runs a small client that connects to a Discord-run backend. The backend hands out one task at a time from fair, per-worker queues, so nobody is starved and the work spreads evenly. When a worker finishes, they report back, and the result is counted against them. A background reaper takes back any task that has been sitting too long and gives it to someone else.

• One-task-at-a-time queue, shared fairly across all workers
• Every finished task is counted against the worker who did it
• A stale-job reaper reclaims and re-queues abandoned work
• Failed tasks get retried up to a limit

The whole point is to see who is working and to pay them right. So the system keeps live stats for every worker - how many they finished, how many failed, their success rate, and whether they are online right now.

There is a full payout flow built for a real team. A leaderboard ranks everyone by work done, and each worker has a profile showing their lifetime totals and earnings. Payouts are logged with a per-task rate, so at any time I can check exactly what a worker is owed and what they have already been paid. Owners and managers get their own powers - managers can run payouts, owners run the whole thing.

• Live per-worker stats - done, failed, success rate, online status
• Leaderboard and per-user profiles with lifetime earnings
• Per-task payout system with a full payment log
• Role-based access - owners, managers and workers

Running a fleet needs control and a clear view. Workers connect with a key that is locked to their own machine, behind an auth token and an IP allow-list, so only my people ever get work. Bad actors can be banned in one click and their tasks are refused at the door.

For the live view, a Discord dashboard updates itself every minute with online counts and totals, and there is a desktop control panel too - a Tkinter app with stat cards and worker tables that refreshes every couple of seconds. It is about 2,700 lines of coordination code, and it turns a messy crowd of workers into a team that runs itself.

• Machine-locked worker keys (HWID) behind an auth token and IP allow-list
• One-click ban or unban, and stat resets
• A self-updating Discord dashboard of the whole fleet
• A desktop Tkinter control panel with live worker tables