I've been working on something for the past few months that I want to talk about.
It started, like most things do, with a problem I couldn't stop running into.
The moment before the plan
I work as a Technical Project Lead at Rimac Technology during the day and run Ñom Studio independently. Between the two, I'm constantly starting things — design systems, client projects, product ideas, the studio itself. And I kept hitting the same wall.
Not motivation. Not time. Not tools.
I didn't know what to do first.
I'd sit down with a fresh project, open Notion or Trello, stare at the empty board, and feel more anxious than before I opened it. Every project management tool I tried assumed I already had a task list. But that was the whole problem — I'd never built this thing before. I didn't know the steps.
I'd end up spending hours researching how other people structured similar projects, cobbling together a plan from blog posts and YouTube videos, second-guessing the order, and eventually just starting somewhere random and hoping for the best.
It worked, mostly. But it was slow and stressful in a way it didn't need to be.
What if the tool figured out the steps for you?
That question wouldn't leave me alone.
The idea was simple: what if you could describe a project in plain language — what you're building, when you need it done, how much time you have per week — and an AI generated an actual plan? Not a vague checklist. A real structure with milestones, prioritised tasks, effort estimates, and clear guidance on each step.
And then, as you worked through it, the plan would adapt. Reject a task that doesn't feel right, and the system re-plans. Finish a milestone, and it recalibrates what's next.
Describe → plan → execute → adapt. That's the loop.
I called it Ruty, short for ruta — Spanish for route or path. Because that's what it gives you: a path through something you've never done before.
Building it
I decided early on that I wanted to build this myself. Not because I have anything against co-founders or agencies, but because the product is deeply personal and I wanted to understand every layer of it.
The stack is React and TypeScript on the frontend, Node.js with Express and PostgreSQL on the backend, and Anthropic's Claude API for the AI layer. Everything is deployed on Vercel and Railway. Authentication is Google OAuth — simple, no passwords to manage.
The AI architecture took the most iteration. Plan generation happens in three stages: first, a conversational intake where Ruty asks clarifying questions about your project (up to eight, depending on how much context it needs). Then it compresses that conversation into a summary and generates the project structure — milestones and tasks with priorities and effort levels. Finally, it fills in the detail fields for every task: what to do, why it matters, how to approach it, and what "done" looks like.
The whole thing streams to the frontend over Server-Sent Events so you can watch your plan take shape in real time. It takes about fifteen seconds, and there's something genuinely satisfying about watching it happen.
The design system shares DNA with Ñom Studio — warm off-white backgrounds, sharp corners, Bebas Neue for display text, Inter for body, and a very deliberate use of red. Red means something in Ruty: it's for actions, active states, and high-priority work. Never decorative.
Where Ruty is right now
The MVP is live and functional. Here's what works today:
You can sign in, describe a project, go through the AI intake conversation, and get a full plan generated with milestones, tasks, priorities, and effort estimates. You can work through tasks, mark them complete, reject ones that don't fit (which triggers the AI to adjust the plan), and track your progress with a points system that reflects meaningful momentum.
There's a progress page with charts and stats, an account page, and a nav system that gets out of your way when you're in the middle of creating something.
Beyond the core product, I've built a custom analytics system (no third-party tracking — everything lives in our own database), a branded email system with weekly AI-generated digests, milestone celebration emails, and gentle re-engagement nudges, and the full payment infrastructure for when we're ready to introduce a Pro tier.
Everything is custom-built. Not because I enjoy extra work, but because I want full control over the experience and I want to keep operating costs as close to zero as possible. The entire stack can serve thousands of users for under a hundred dollars a month.
What's next
The immediate priorities are:
Task validation — when you add a task manually, the AI checks whether it aligns with your project goals and offers to keep it, reject it, or reframe it. This closes the loop between human intuition and AI structure.
Parent/child projects — so you can break a large project into sub-projects, each with their own milestones and tasks, while tracking progress at the top level.
Gamification — badges, streaks, and levels that make progress feel tangible. Ruty already has a points system, but there's room to make it more rewarding without making it gimmicky.
Integrations — Notion and Google Calendar are the first two. Syncing tasks to a calendar and linking project context to existing Notion workspaces are the most requested features from early users.
Mobile — Ruty is responsive today, but a native app is on the horizon. The kind of tool that helps you figure out what to do next should be in your pocket.
The bigger picture
Here's what I keep coming back to.
The productivity tool market is enormous, and almost every product in it is built for people who already know how to manage projects. Notion is incredible — if you know how to set it up. Asana is powerful — if you already have a workflow. Linear is beautiful — if you're running a dev team.
But there's a huge group of people who never adopt any of these tools. Not because the tools are bad, but because they require knowledge the user doesn't have yet. They need to know their tasks before they can manage them. That's a prerequisite that nobody talks about.
Ruty is for the moment before the plan exists.
It's for the freelance designer launching their first product. The solo founder who has a vision but no roadmap. The creative who knows what they want to make but not how to break it into steps. These people are already anxious about not knowing where to start — the last thing they need is a tool that makes them feel more behind.
Ruty is meant to meet them there. Not as a sharp tool that happens to be friendly, but as a warm companion that's also sharp.
Why I'm writing this on the Ñom Studio blog
Ruty is a sister product of Ñom Studio. They share visual DNA, they share a philosophy about making quality tools for independent creators, and honestly, Ruty exists because I needed it to build the studio itself.
If you're the kind of person who reads this and thinks "I need that," you're exactly who it's for. Ruty is in beta right now and I'm actively building based on user feedback.
Try it, break it, tell me everything.
— Raul