I finally launched Xeyria — trying to solve the “I forgot what I was doing” problem in software projects
Today I launched Xeyria on Product Hunt.
The idea came from a problem I kept experiencing while working on software projects: after stepping away for a while, getting back into the project often meant reconstructing context — what I was working on, why certain decisions were made, what changed, and what was still unfinished.
AI coding tools are great at helping write code, but I wanted to explore a different problem: how do we preserve useful engineering context between sessions?
So I built Xeyria.
It preserves things like codebase context, architecture, technical decisions and discussions so you can pick up where you left off.
Today is the first public launch.
I'm much more interested in learning from other builders than collecting vanity numbers, so I'd love to hear:
How do you currently get back into a project after being away from it for a few days or weeks?
One thing first, because it's your launch day and it's still fixable. The paragraph starting "This fits Indie Hackers much better because..." reads like notes that came along rather than something meant for the post.
To your actual question. I keep one file per project holding only the ACTIVE state, and a separate archive for the history. The split does more work than the format does.
But here's the failure mode I'd flag, because it's the one you're building against. Today I opened my own status file, found three open tasks, and started on them. They had been finished two days earlier. The file that said so was sitting in the same folder. The "why" wasn't lost — it was there, it was outdated, and it looked exactly like the parts that were still true.
So storing the reasoning is the easier half. The hard half is that a note which was right when written and is wrong now looks identical to one that still holds. Nobody re-reads a note to check. If Xeyria can make an expired decision look different from a live one, that's the part I'd pay for.
And a trade, if you want it. I spend 30 minutes on your landing, you spend 30 on my entry page. Not opinions — a task. I'd hand you something like "find out what this costs and how long it takes", then shut up and watch where you stall. You'd do the same to me. Mine is a five-question quiz page for a product I sell. It's the part I already suspect is broken. 92% of people who land there never click a single question, and the numbers can't tell me whether they left or just sat there. No hard feelings if it's not your thing, and no follow-up from me.
This is a really good callout, especially the part about outdated information looking exactly like current information. That's a problem I care about a lot with Xeyria — context isn't useful if you can't tell whether it's still relevant.
And yes, I'm up for the trade. I like the “task, not opinions” approach.
Send me the entry page and the task you want me to do. I'll spend 30 minutes going through it and tell you exactly where I got stuck, what felt unclear, and what I'd change.
Here you go: mvpbuilder.io/go?utm_campaign=swap&utm_content=swap01
The task: find out what it costs and how long it takes. Nothing else. Don't read it as a reviewer — just try to answer those two questions, and stop at the point where you can't.
The useful part for me is where you stall, backtrack, or guess. Not whether it reads well. If you get annoyed and close the tab, that's the most valuable thing you can send me — just tell me which screen it was.
I won't explain anything while you're in there, and I won't argue with what you find. That's the whole deal.
Your half: what do you want me to do on Xeyria, and where do I start? Same rule — hand me a task, not "what do you think", and I'll tell you exactly where I got stuck.
Perfect. Your task:
Start from xeyria.com and find out how to get Xeyria working in VS Code for your first project.
Nothing else — don't review the product or tell me what you think. Just try to get from the landing page to the point where Xeyria is actually working in VS Code.
Stop wherever you get stuck, confused, or have to guess. Tell me the exact screen/step where that happens.
Done — 18 minutes, and I never reached step 2. So this is where I stopped, not where I finished.
Started on the landing page, clicked "How It Works".
"Get started in under a minute" set my expectation before anything else did. I'd read that as under a minute for someone who installs VSIX packages regularly.
Step 01: "Download the VSIX package. Install Xeyria inside VS Code." I didn't know how to do either one. I left your site to look it up. That's the first place I stopped.
Back in VS Code I searched the Marketplace for "vsix", because that was the word I'd been given. I got VSIX Viewer, VSIX Manager, Pack VSIX, VSIX Installer. I concluded that VSIX was a thing I had to install first.
So I installed "VSIX Package Inspector" — 740 installs, not yours. Its context menu has "Download VSIX", which confirmed the wrong model I'd just built: install first, download after.
Fourteen minutes in, I searched the Marketplace for "xeyria" on a hunch. It was right there, one button, Install. Everything from point 2 onward had been a detour.
I clicked Install and got "Do you trust the publisher Xeyria? Xeyria is not verified." I clicked through because I was testing. I couldn't tell from anything I had seen whether I should have. Same question I'd had about the VSIX file earlier: nothing I read told me how to check that either one was actually yours.
I stopped there. Never got to Authenticate.
This is exactly the kind of test I was hoping for. Thank you for documenting the actual path rather than just giving me an opinion.
The VSIX wording sending you into a 14-minute detour is especially useful, and the publisher trust question is something I hadn't seen from a first-time user's perspective.
And you're right about the distinction between where you stopped and where you finished — that's the important part here.
I'll take this back as an onboarding problem to investigate rather than trying to explain it away.
You named the useful part yourself: where someone stopped, not where they finished. That's the one thing I'm still missing from my side — which screen on my quiz page you stopped at, or had to guess. One line, no reasoning, no polish. I won't argue with it. And if it's faded or you'd rather not, just leave it. I won't ask again, and the trade was worth it either way.
Congrats on the launch. My honest answer to your question: I keep a plain markdown file per project — what I was doing, what's next, open questions. Low-tech, but it survives everything. The pain is real though; I lose the "why" behind decisions way faster than the "what." One thing I'd watch with Xeyria: make sure it captures the why, not just the state. State is easy to reconstruct from code. The reasoning behind decisions is the part nobody recovers well, and that's where the value is.
That's a really good distinction. I agree — the “what” is usually recoverable from the code, but the “why” is what gets lost.
That's actually a big part of what I wanted Xeyria to preserve: decisions, discussions, and the reasoning around changes, not just a snapshot of the codebase.
The markdown file approach makes a lot of sense too — low-tech and durable. I'm curious: do you usually update it as you work, or mainly when you're about to step away from a project?