Why I Built Arcapush
By Mojeeb Titilayo — Founder, Arcapush | @MojeebMotion | mojeeb.xyz
There is a problem I keep meeting no matter what I build:
shipping is only half the work. Being found is the other half.
AI has made software creation faster and more accessible. It has not made attention infinite. In fact, when more people can ship, discovery becomes more competitive.
Arcapush exists because I wanted infrastructure for that second problem.
I learned by building before I had a title for it
I grew up around Iware and Fiditi in Nigeria. My earliest relationship with computers was curiosity before credentials: school computers, PowerPoint, forums, blogs, copied snippets, JavaScript experiments, WordPress, and years of learning by making things work.
I studied History at the University of Ilorin, not Computer Science. That did not stop the building habit. Over time the work moved through content, marketing, Web3 communities, product strategy, and eventually AI-native software building.
Under BlindspotLab, I kept shipping products across different problems and formats. The exact product count matters less than the pattern: build something, put it in public, learn what gets attention, learn what disappears, repeat.
That repetition is what led to Arcapush.
The problem was not that people stopped building
The opposite happened.
AI-first development tools lowered the time required to move from an idea to a working product. Solo founders, designers, marketers, researchers, domain experts, and engineers could all create more software with smaller teams.
But the distribution layer did not become easier at the same rate.
A good product could still disappear because the founder had no audience, no launch machine, no repeatable distribution workflow, or simply moved on to building the next thing.
That is the gap Arcapush is designed around.
What Arcapush became
The first versions of Arcapush were narrower and more focused on vibe-coded startups. That description is now outdated.
Today Arcapush helps products, AI agents, and hackathon projects get discovered, understood, and distributed.
The operating loop is straightforward:
DISCOVER → VISIT → TRACK → DISTRIBUTE → MEASURE → LEARN → DISTRIBUTE BETTER
That means Arcapush is not just a list of links. The platform now has several connected layers:
- ◆a global discovery surface,
- ◆structured project records,
- ◆project media and richer editing,
- ◆builder identity and ownership verification,
- ◆founder analytics,
- ◆managed distribution inquiries,
- ◆a public CLI,
- ◆a permanent global Genesis registry,
- ◆a dedicated canonical surface for AI agents at agents.arcapush.com.
Products and hackathon builds remain part of the primary Arcapush application.
Why Genesis matters to me
My History background shows up most clearly in the Genesis Registry.
Every Product, AI Agent, and Hackathon build receives a number from one global sequence. That number is provenance: a record of when the project entered Arcapush's history. It does not change because a founder edits the page, claims ownership, verifies an identity, or changes positioning.
The registry can contain gaps because database sequences are not recycled and non-public records can retain their issued numbers. That is intentional. The point is not to manufacture a perfectly continuous leaderboard. The point is to preserve the record.
That is closer to how I think about documentation: what happened, in what order, and what should remain true later?
What Arcapush is not anymore
An older version of this founder story described USDC payments, SOL payments, Paystack checkout, paid pinning, and an on-chain boost contract as if those were still core Arcapush features.
They are not.
Arcapush removed public self-serve checkout and paid pinning. Discovery, normal submission, and founder analytics on owned listings are free. Promotion is selected, managed distribution work through arcapush.com/promote, covering areas such as launch strategy, product positioning, content distribution, founder storytelling, community activation, and ecosystem campaigns.
There is no public checkout.
That change matters because the product became clearer: Arcapush should not charge founders simply to exist in the registry. If we do paid work, it should be deliberate distribution work with an actual service attached.
The build process is still part of the story
Arcapush itself is AI-assisted software. I use modern coding agents and models heavily, but the useful lesson is not which model generated which line.
The work is in deciding what the product should become, giving agents enough context to implement safely, reviewing what they change, testing the result, and continuing when the first implementation is wrong.
The current application is built with Next.js, TypeScript, Prisma, PostgreSQL, NextAuth, and Vercel, with Brevo handling email delivery. The public Arcapush CLI lives separately at github.com/mojeebdev/arcapush-cli and is published as @blindspotlab/arcapush.
That stack will change. The product principle matters more.
What I want Arcapush to become
Software is becoming infinite. Attention is not.
I want Arcapush to become infrastructure that helps strong independent work survive that imbalance: a place where a project can have a durable public record, a builder can prove connection to it, an agent can have a clean canonical identity, and distribution can be measured instead of treated like magic.
There is still a lot to build. The value of writing this founder story again in August 2026 is that Arcapush has already changed enough to make the March version inaccurate.
That is a good problem. It means the product is moving.
If you built something worth discovering, submit it.
If you want to explore the registry, start at arcapush.com/registry.
If you want managed distribution help, use arcapush.com/promote.
We're here to make sure good work has a better chance of being found.
Update note
This founder story was materially revised on August 20, 2026 to match the current Arcapush product. References to removed public checkout, paid pinning, crypto payment flows, outdated framework versions, and unsupported investor-discovery claims were removed.
