At some point in year two of doing performance marketing, I noticed something. Every problem I was solving for clients — bad reporting, shallow research, wasted time on data collection — had the same shape. Someone was spending hours gathering information so they could spend thirty minutes acting on it. The ratio was wrong. It was always wrong.
I didn't set out to build a suite of tools. The V Projekt grew out of fixing specific things that annoyed me, one by one. What surprised me is that they all ended up solving the same underlying problem, just in different contexts.
Here's everything under The V Projekt — what each piece does, where it stands, and what the whole thing is actually about.
What is The V Projekt?
The V Projekt is a personal lab — a loose holding structure for tools and brands I build alongside client work. Not a startup. Not an agency. Something closer to a one-person R&D operation where the research is "what breaks in my daily work" and the development is "build the thing that fixes it."
The name is simple. It's mine. Every project under it is something I built from scratch, without a co-founder, without a team, usually without a budget. The common thread across all of them is automation — specifically, automating the information layer so the person who needs to act can skip straight to acting.
Why build tools instead of just using existing ones?
Honest answer: because the existing ones didn't do the exact thing I needed. There are investor research tools, but none that generate a per-partner brief in ten minutes from a name and a fund. There are domain research tools, but none that run the full filter sequence — spam score, Wayback history, authority threshold, availability — automatically, on a schedule, without you touching them.
The deeper answer is that building forces you to understand the problem at a level that using never does. When you automate something, you have to decompose every step. That process of decomposition has made me better at my day job in ways I didn't expect.
"Every tool I've built started as a process I was doing manually and then got tired of doing manually."
There's also a practical economics argument. The tools I've built have either saved me more time than they cost to build, or they've become products that other people pay to use. PitchEdge hit production and now has over 200 founders using per-investor intelligence briefs before their meetings. Domain Radar checks 10,000 expired domains weekly and flags the 0.2% worth acting on. The payoff is real.
What does each project do?
PitchEdge — per-investor intelligence for founders
The problem: founders research the fund. They should research the partner.
There's a consistent gap in how founders prepare for investor meetings. They read the fund's website. They update their deck. They miss the three things that actually determine whether a meeting advances: the partner's real investment thesis (not the public one), any portfolio conflict that would quietly kill the deal, and the specific objection that partner always raises at seed stage.
PitchEdge fixes this. You give it a name and a fund. It generates a brief covering all three — sourced from investment records, the partner's writing and speaking history, and portfolio data — in a format you can read the night before. 200+ founders are using it ahead of partner meetings.
Visit PitchEdge ↗Domain Radar — automated expired domain research
The problem: expired domain research is six hours of repetitive work to find three good candidates.
99% of expired domains are junk. The 0.2% that aren't require checking spam score, Wayback Machine history, domain authority, referring domain count, and current availability — in sequence, for every domain on the list. I spent two weekends doing this manually before I automated it.
Domain Radar runs on n8n, pulling fresh expiry lists weekly, running each domain through five filter stages, and outputting a clean shortlist with the full data snapshot for each candidate. The pass rate is 0.2%. That's not a failure — that's the point. The whole value is that you never see the 99.8% that don't qualify.
Read the full build story →Declaro AI — reporting that writes itself
The problem: the person best positioned to interpret campaign performance is spending five hours copying numbers into a Google Doc.
Agency reporting has a structural problem. The work that requires human judgment — "what changed, why, and what should we do differently next week" — comes last, after hours of data entry that require no judgment at all. By the time someone gets to the thinking part, they're exhausted by the copying part.
Declaro pulls live Meta and Google Ads data via API, feeds it into Claude with client-specific context, generates a structured narrative report, and delivers it to an Airtable base for one-click client sign-off. The account manager reviews and approves. They don't write. Three to five hours becomes thirty minutes, and the thirty minutes is the part that actually requires them.
What is Jarvis?
Jarvis is different from the others. The other tools solve external problems — things clients and users face. Jarvis solves an internal one: how do you manage the cognitive load of running multiple projects, client work, and a personal life without constantly operating in reactive mode?
The name is intentional. Tony Stark didn't have a team — he had a system that watched over his operation, told him what needed attention, and deployed specialized capabilities when the situation called for them. That's the exact model I'm building toward.
In practice, Jarvis is a multi-agent system that monitors my work environment, surfaces priority tasks based on context (what's urgent, what's been neglected, what has a deadline approaching), and spins up specialized sub-agents for different types of work — research, drafting, analysis, scheduling. It's a prototype right now. But the vision is a personal AI OS: not an app you open, but a system that runs alongside everything else and keeps the operation coherent.
Jarvis is deeply personal — it's built around my specific workflows, my specific context, my specific definition of "priority." Productizing it would mean abstracting that personal fit away. For now, it's a tool I'm building for one user: me. If it proves out at that scale, the product question becomes interesting.
What about Harit?
Harit is the odd one out in The V Projekt. It's not an AI tool. It's a premium microgreens brand — the name comes from a Sanskrit word for green, and the positioning is B2B wholesale, India to Germany. I built the full brand: identity, site architecture, SEO, content strategy, wholesale positioning.
I include it here because the building process was the same as the tools. Start with a specific gap. Understand the market deeply. Build the infrastructure that lets the business operate — in this case, a site that's indexed, has a real brand story, and can convert wholesale inquiries without me explaining the concept from scratch every time.
The skill that made PitchEdge good at explaining a complex product clearly is the same skill that made Harit's brand voice work. The tools and the brands train the same muscle. I've found that going back and forth between them keeps the thinking sharper than staying in one domain.
What's the connecting thread?
Every V Projekt tool automates the information and research layer — the data collection, the fact-finding, the synthesis work that happens before a human can make a useful decision. PitchEdge automates investor research. Domain Radar automates domain research. Declaro automates performance data collection and narrative generation. Jarvis automates the meta-task of knowing what to work on next.
The human's job, in every case, is to act on the output. Not generate it.
This isn't an accident. It comes from spending years doing performance marketing — a field where a surprising amount of work is assembling information before you can do the actual thinking. The pattern repeats across almost every knowledge-work role: analyst, founder, marketer, recruiter. Someone spends 70% of their time gathering data so they can spend 30% making decisions.
Flip that ratio and something changes. The decisions get better because they're made with more attention. The work gets faster because the gathering is automated. And the person doing the work stops dreading the repetitive parts because the repetitive parts don't exist anymore.
That's the V Projekt thesis. It's simple enough to fit in a sentence but interesting enough to keep generating new applications. I don't think I'm close to running out of problems to solve.
What's next for The V Projekt?
SOMA is next. It's an Ayurvedic wellness brand — premium positioning, German market, B2B wholesale like Harit. The brand identity work is underway.
The V Projekt itself is about to get louder. Content, builds in public, the whole thing. Until now it's existed mostly as a label for what I was already doing. That changes soon.
If you're building something and the research or data collection is the part eating your time, the tools above are worth looking at. And if the problem you're trying to solve sounds like something I'd build — get in touch. Most of The V Projekt started as a conversation about a problem someone had that they'd just accepted as part of the job.