The founder's note
I failed about twenty PCBs.
So I'm rebuilding the whole thing.
You can use the actual AI before you pay us a cent, because if it's any good you'll know in five minutes, and if it's trash you shouldn't be paying anyway.
Rishik · ECE @ UT Austin · building ProtoFlow
attempt #21ch.01/the deal
Try the AI before you pay.
We're the only PCB tool that lets you.
none of that here
Every other tool makes you earn it first. You start a trial, the clock starts ticking, they grab your card, and you burn through your free credits in one afternoon and hit a wall right when it starts getting good. I hated that, so I just didn't do it. That's honestly the whole pricing pitch.
ch.02/what this actually is
It's not an AI sidebar.
what it's not
A chat bubble stapled onto the corner of someone else's editor so the box can say "AI" on it.
what it is
A from-scratch PCB tool built around how hobbyists, students, and engineers actually work.
I'm not trying to decorate the old way of doing things, I'm trying to replace it. The goal is to change how you build hardware, not bolt an assistant onto the same editor that's been frustrating you for years. It's a big swing, and I'd rather take it than ship another sidebar.
ch.03/the real problem
The hard part was never the circuit.
It's the thousand tiny chores standing between you and a finished board, and every one of them is just annoying enough to break your focus. Here's what sourcing one part actually looks like, in real time:
T+0:00
Find a module. This Thorlabs one looks good. Now I just need the datasheet, which somehow takes ten minutes of digging to track down.
T+4:30
Next up, the matching sensor. The distributor doesn't have it. The manufacturer's site doesn't even have a symbol for it. Nobody has it.
T+9:00
Fine, I'll draw the symbol myself. From scratch.
T+13:00
And the footprint too, where every single pad has to line up perfectly or the board comes back from the fab useless.
T+20:00
One part done. Cool. Fifteen more to go.
That's not engineering. That's data entry with extra steps.
ch.04/exhibit a: the rap sheet
A short history of failing.
High school
A board to turn my lights on and off from my phone.
Also high school
A board to put an alarm inside my pillow so I would actually wake up.
Going simpler
Just a basic ESP32 breakout. How hard could it be?
~20 boards later
You'd think I'd take the hint and go get a normal job.
ch.05/the last straw
The board that finally broke me.
I'm an ECE student at UT Austin, and I got brought onto a project building a PCB that helps doctors measure blood flow in the brain. We're talking a high-speed 1MHz DAC board driving lasers, with actual patients on the other end of it. So after twenty boards that did basically nothing, the first one that really mattered was a medical device. Great.
And it was every chore from the grind above turned up to eleven. I'd hunt down a laser driver, dig up its datasheet, then go looking for a sensor that apparently doesn't exist anywhere on the internet, draw the symbol myself, draw the footprint myself, line everything up, and then do the whole thing again for the next part. In between I'm bouncing from the datasheet to KiCad and back, pasting a block into ChatGPT to ask how it works, realizing I forgot to give it half the circuit, and slowly losing my mind.
"Why is PCB design so hard?"
me, at 2 a.m., to absolutely nobody
Enough was enough. Somewhere in that mess I stopped being mad at myself and got mad at the tools instead, because seriously, what if someone built a PCB editor around how hobbyists and students and engineers actually work now, instead of how somebody decided we should work decades ago? That's where ProtoFlow started. It wasn't a business plan, it was a bad night with a laser board.
ch.06/so i built it
So I built the editor I wish I'd had.
This is what circuit design looks like in my head when the tools aren't fighting me.
One search, every vendor
Search every distributor in one window and open the datasheet right there, natively, without leaving the app.
Split-screen the datasheet
Pin the datasheet next to your schematic so you never have to alt-tab to KiCad again.
Tabs
Jump between projects and views like you would in any real editor.
Panels that move
Drag the workspace into whatever layout you want instead of whatever a menu decided for you.
Live Git, visual diffs
See exactly what changed in your design, live, so you stop guessing what you broke.
Real-time collaboration
Edit the same design together live, or lock people to comment-and-view when you only want feedback.
Comments on a schematic
Leave comments right on the design, exactly where the problem is.
A community
Browse everyone's designs in one place, then clone them, fork them, and upvote the good ones.
Native KiCad files
It reads and writes the real KiCad format, so you're never locked in.
Built-in AI
It finds parts, places them, and wires up the boring stuff. Sometimes it faceplants, but it still beats a blank schematic.
And there's a whole PCB layout side I haven't even gotten into here. I'll just let you try it.
ch.07/one honest disclaimer
It might be amazing. It might be trash.
You're the one who gets to decide that, which is the entire reason you can try the AI before paying for it. And just so we're clear, none of this is me dunking on KiCad. KiCad is great, I love it, I still use it, and ProtoFlow reads and writes native KiCad files for exactly that reason. If you ever want to go back, your files are sitting right there waiting for you.
epilogue / your turn
Go break it.
Download it, throw a real board at it before you pay a cent, and come tell me what you actually think.