My guitar teacher, Fábio, pointed me to the BOSS GT-1, a 2016 multi-effects pedal that simulates amps and dozens of effects. It does almost everything. The hard part is getting there: hundreds of settings spread across menus named after gear.
I spend my days designing interfaces, so I got the logic quickly. I also saw who it was made for: people who already think like a sound engineer. If you just want to play a song, you usually stay on the factory sound.
I didn't want to learn the name of every knob. I wanted to ask for the sound.
This is the first of two pieces about GT Pilot, the tool that came out of it. Here I cover the path from a personal annoyance to the first signs that the problem is real for other people and that the idea is well received. The second one will cover what comes after interest: how people get to use it, and whether they come back.
The first two days
The official BOSS app keeps a map of every pedal setting in a file. With that map, Claude could read and change any parameter over the USB cable.
I asked for a tone, it sent the commands, I played and said nothing had changed. It had. I checked the messages going to the pedal and they were identical to the official app's, but the change was too subtle for someone expecting to hear a difference. Even with everything working, I had started to distrust the tool. So the first change became big enough to hear, and the patch name started showing up on the pedal's display right away.
On the second day, I asked it to save a tone and Claude wrote over the Iron Man patch I had built the day before. I had a backup and got it back, but it was clear I couldn't rely on the model being careful. Saving now has three steps the code won't let you skip: show what's in the slot, wait for a yes, and keep a copy of what was there. Volume got a similar rule after a request came out far too loud, and today no setting moves more than 15 points at a time.
None of my requests used technical names. They were things like "I want the Knee Socks solo", "needs more drive", "it sounds weird" and "turn it down, neighbors". The tool's job is to turn that into a small adjustment and say what changed. When I said "it's noisy", it touched only the noise suppressor and left the rest of the tone as I had approved it.
Two weeks later, I tested it with Fábio. We built three tones by talking: Comfortably Numb, Sunday Bloody Sunday and Rosanna. On the U2 one he found the sound too clean, Claude raised the amp gain, and we got there.
Up to that point, it was a tool that worked for two people in the same room. The next question was whether the problem existed outside it.
The problem isn't just mine
Before sharing it, I looked for people already trying to solve this, and found plenty.
On a big guitar forum, someone asked how to get close to the Stairway to Heaven solo tone with a Les Paul, a Marshall and a GT-1. It was the same request I was making to Claude, written by someone who had never heard of me.
There are sites with more than a hundred hand-built GT-1 patches, organized by song and artist, and channels selling packs of ready-made tones. People dig around and pay to get the sound of a song.
The maker of another pedal launched a feature this year built on the same idea: you describe the tone and the preset goes straight to the device. I also found an open project that generates preset files for the GT-1's successor, a phone app in testing that controls the GT-1, and a guitarist who commented on my post that they're building something similar.
I liked not being the first. Several people attacking the same problem from different angles is a sign it's real. What GT Pilot does differently is adjust the pedal while you play, with a backup and undo, instead of generating a file to import later.
Showing it to people who own the pedal
In mid-October, I posted in four Facebook groups for GT-1 owners, in Portuguese and English, looking for testers.
In the largest one, in English, the first reaction was a question: "Is this real?". I replied that I'd built it for my own GT-1 and that it was free, and the conversation changed from there. A software engineer wrote that the idea was amazing and suggested a web version. Someone else commented "Impressive". In the Brazilian group, someone replied they'd test it because they had the pedal. The English post got 18 reactions and 13 comments, in a group that now sees about one post a month.
Then I ran small ads on Facebook and Instagram for about ten days, on a test budget. They reached a little over ten thousand impressions, a few hundred clicks and around thirty people leaving their email. The ad that brought in the most people didn't mention any feature. It just said "I want the Stairway solo tone". The ads by music style and the version aimed outside Brazil did worse.
What people asked for
The most useful part came without me asking. The requests weren't for new features. They were for the same product, in each person's own context.
More than half of the sign-ups asked for Windows. The comments asked for mobile, iOS, Android and web. The sign-up form and emails brought up other BOSS pedals, like the GT-1000, GT-100 and Katana. Some people asked to use it with ChatGPT.
I treated those requests as a work queue. The Windows test version came first. The first tester sent a screenshot of an error, the fix went out the same day, and later came word that the pedal was recognized. Another tester, on Windows 11, wrote that it worked very well and that they were surprised by how much it helped them set things up. Next came the ChatGPT plugin and early support for the GT-100, which is still waiting for a unit to test on.
Where I am
That's close to 60 sign-ups and more than 80 downloads in three weeks. The problem is real for other people, and GT-1 owners get the idea quickly and want it on their own computer and their own pedal.
That validates interest, but not the product yet. Most people arrive on their phone, and the tool needs a computer, a cable, the BOSS app installed and an AI subscription. The usage I can measure is still low, and I don't know how many people come back for a second session.
Part 2 is about that: the path from the click to the second use, and what I learn once I start watching people use it.
