Full disclosure
How GigCharts is made
An AI assistant wrote most of the code behind GigCharts, and helped write most of the words on this site. I'd rather tell you that plainly than have you wonder.
On this page
This isn't a confession and it isn't a boast. It's just how this software got made, written down, so you can decide what you think about it.
The app
GigCharts is the fourth version of a tool I first built in 2002. The idea, the layout rules, the opinions about what a good chart looks like on a music stand at nine in the evening — those are mine, and they took twenty years of playing to arrive at.
Most of the code in this version was written by an AI assistant — Claude — working from my instructions, inside an architecture I laid out, one small piece at a time. I read what comes back, I run it, and I use it on my own gig charts before anyone else sees it. Plenty of it goes back for another pass, and some of it I throw away.
So the work is human-directed and machine-written. What doesn't move is the responsibility: if a chart prints wrong, that's mine to fix, and I'm the one who answers the email.
This website
Same arrangement for the words. I decide what a page should say and what's true in it; a language model helps me draft and tighten; I edit every line that ships. If a sentence here sounds like me, it's because I kept it.
The reference pages are a bit different. The format spec, the import tables, the ChordPro coverage list and the element reference are generated from the app's own data rather than from anyone's memory of it — so when the code changes, the docs change with it.
The pictures
Every screenshot on this site is a real screenshot. They're captured by a script that drives the actual app and exports the actual PDF — no mockups, no retouching, nothing drawn to look better than the thing does. The demo video on the front page is a recording of that same app, at that same speed.
A few of the decorative images were generated with ChatGPT. Where you're looking at artwork rather than the product, assume a machine had a hand in it; where you're looking at the product, you're looking at the product.
Your songs
No AI touches your songs. GigCharts lays out, renders and exports entirely in your browser. Your charts are never sent to a model, and nothing you write here is used to train anything.
The one place AI appears in the product is optional and you start it: Make a chart with AI is a spec you hand to your own assistant, which writes a song file that you then open in the app. GigCharts never calls a model on your behalf.
If you sign in and sync, your songs go to my server so your own devices can share them — that, and only that, is in the privacy policy.
Why I work this way
I care a great deal that music is made by people. Playing it, singing it, arranging it, deciding that this verse should drop to a whisper — that part is ours, and I have no interest in a machine doing it.
Building the tools is different work. This version exists because the help made it possible for one person, with a day job and a band, to finish something that three earlier attempts across two decades never quite reached. Written any other way, the honest answer is that it wouldn't exist at all.
Made by hand where it counts, with help where it doesn't. If that's not for you, I understand — better you know it here than guess at it later. Either way, the chart you print is yours, and nothing in it was written by a machine unless you asked for it.