Solomon
Huckstep

A printer-agnostic print spooler built from scratch. A REST API and Blazor Server dashboard that queues print jobs, delivers them to network printers over IPP, pushes live status to the browser with SignalR, and logs a full audit trail of every job. Printers are discovered over mDNS and picked up at runtime.

Each printer gets its own monitor with its own cancellation token, so one stalled printer never blocks the others. After a crash, unfinished jobs are re-checked against the printer and marked failed if completion can't be confirmed. Verified end to end against a physical printer.

A terminal application built for Hachette Book Group, releasing carton label print jobs in its distribution warehouse and watching them through to completion. It splits the release file, distributes the pieces across five label-rendering instances, tracks every file with folder watchers, retries failures, and can be cancelled and rolled back cleanly mid-run.

A C# rebuild of an earlier PowerShell script written by an IT staffer. I talked through the original logic with its creator and read the script myself to confirm the new version's output matched before building on it, keeping the same workflow while adding more readable code, a clearer live monitor built with the Spectre.Console library, and a Teams summary posted after each run as a record that it ran.

Built for Hachette Book Group, generating shipping labels roughly 300 times faster than BarTender, the commercial software we used, measured on the same production hardware. Builds printable label pages from order data and sends PCL directly to printers over raw TCP on port 9100. Templates are JSON, with layouts positioned on a normalized 0 to 100 scale, and pre-rendered elements are cached so they aren't regenerated. Built for the 4-up layout. Rewritten from PowerShell to C#, roughly doubling throughput.

Demoed as a proof of concept for print-on-demand. A 15,000-label release that took about 40 minutes through BarTender would render in about 10 seconds, so labels could be generated when the floor needs them instead of pre-generated in bulk.

While building it, I found the format's documentation didn't say which of the 100 data fields each label type actually needed. I started a tool to analyze a year of real files and classify each field as required, often used, optional, or never used, so a production version could validate input properly.

Running four industrial printers already took the print operator's full attention, watching for jams, page misalignment, and toner that wasn't curing properly. Operators also had to find files by hand across several waves and label types at once, reprint by Ctrl-F'ing to a line range without breaking the label order, and hand-log paper counts on paper, then again into a spreadsheet.

PrintFlow is a terminal app for monitoring, queuing, lookups, reprinting, and logging label print jobs. Files are picked by number, supporting single, list, range, or wildcard input, and queued to multiple printers at once. Reprints go by header or by label range, with the order auto-corrected, and it logs paper counts, user actions, and which printer each file went to.

Multiple print operators used it daily. It replaced several disconnected manual steps with one tool and let them focus more on the job they're actually paid to do.