Viktor has become a core part of how we run operations at UniTru, giving us leverage that would normally require engineering resources. It allows us to manage workflows across Slack and Google Sheets seamlessly, handling things like scheduling, status tracking, and coordination in real time. What stands out most is how quickly we can build, test, and refine workflows without slowing down, which is critical for a fast moving team. At a high level, Viktor does not just save time, it enables us to operate at a level of speed, organization, and scale that would not be possible otherwise.
getviktor.com
Paweł here, Chief of Staff at Zeta Labs - the team behind Viktor.
We use Viktor internally every day. The difference before and after installing a skill is not subtle. Without one, Viktor pulls your data and summarizes it. With the Meta Ads skill installed, it catches learning phase disruptions, flags audience overlap, identifies creative fatigue patterns - things a senior media buyer spots in 30 seconds but take hours to diagnose from raw data.
Skills turn Viktor from a capable generalist into someone who actually knows the domain. That jump in quality is what convinced us this was the right next step.
getviktor.com
Hey everyone, Fryd here. Co-founder of Viktor.
This is our third Product Hunt launch. The first one hit #4 for the day. The second went vertical into media buying. Both taught us the same thing: Viktor's value depends entirely on what it knows about your specific domain. So we built a way for anyone to teach it.
We shipped the Skills Marketplace.
The idea: anyone with deep domain expertise can package their playbooks into a skill that changes how Viktor reasons about a specific domain. Think Claude Code skills, but for business operations. We're starting with ads. More verticals coming.
Skills are expert-built knowledge modules. Viktor already connects to your tools and executes actions. Skills teach it how the best operators actually use those tools. Without one, you ask Viktor why your CPA spiked and it pulls the data. With one, it applies the same logic a senior performance marketer would. It knows what to look for, what to ignore, and what to escalate. Different quality of answer from the same question.
The first two skills:
-> A Google Ads skill from Julio, an agency founder who manages campaigns across multiple clients. How he catches waste, structures accounts, and scales what's working. The patterns he's built managing real budgets across dozens of accounts.
-> A Meta Ads skill from the Head of Growth at Wispr Flow. The actual diagnostic framework he runs on accounts with real spend. Not theory. The exact logic he uses to read account data and decide what to fix first.
We started with ads because the feedback loop is fast. ROAS goes up or it doesn't. No hand-waving. But the marketplace is built to be general. We want skills for finance, ops, recruiting, support. Anywhere someone has spent years developing judgment that can be codified.
One expert packages their playbook once. Every team that installs it gets that expertise applied to their accounts, every day, without hiring that person.
If you have deep expertise in any vertical and want to package it into a skill, we want to talk.
@fwiatrowski third launch congrats fryd. the marketplace concept is smart most experts have playbooks but no way to scale them without hiring more people. curious how you’re vetting the skills for the marketplace?
getviktor.com
@priya_kushwaha1 thanks! right now it's very hands-on. we're not opening a self-serve upload and hoping for the best.
every skill in the marketplace comes from someone running real campaigns with real spend. Julio manages Google Ads across multiple agency clients. the Meta Ads skill comes from the Head of Growth at Wispr Flow - actual diagnostic framework he uses on live accounts.
we picked ads as the first vertical specifically because vetting is binary. ROAS goes up or it doesn't. CPA improves or it doesn't. there's no way to hide behind vibes when the numbers are right there.
the process: we work directly with the expert, turn their playbook into a skill, then test it against real account data before it goes live. if the logic doesn't hold up, we don't ship it.
as we scale the marketplace we'll build more formal review layers - but starting curated and small is the whole point. quality over quantity.
getviktor.com
Peter here - Technical Co-Founder of Viktor.
The engineering question behind Skills: how do you make an AI reason like a domain expert without retraining it?
Our answer is modular knowledge injection. Each skill is a distilled expert playbook - decision trees, benchmarks, diagnostic frameworks - that plugs into Viktor's reasoning at inference time. A question like "why is my CPA up 40%" triggers a full diagnostic chain: data pulls, framework application, benchmark comparison, prioritized actions. All from one prompt.
We started with ads because the feedback loop is immediate. ROAS goes up or it doesn't. The architecture is domain-agnostic though - if you have deep expertise in any vertical, the same skill system works. Excited to see what the community builds.
getviktor.com
Bojan from the Viktor engineering team here.
Each skill injects structured domain knowledge into Viktor's reasoning without breaking general capabilities. Not a prompt template - modular knowledge that composes with everything Viktor already knows.
We spent real time making sure installing a skill adds an expert without losing the generalist. Curious what external builders will create with this.
getviktor.com
Bartek here ✌️I designed the Skills Marketplace.
Under the hood, skills are complex - decision frameworks, benchmarks, diagnostic logic. For the user it needed to feel as simple as adding an app. In fact, we wanted to make you feel like Neo in Matrix downloading expert knowledge to Viktor's brain brain, one click away.
This is just a start. We're open to feedback on how to make it even better.
getviktor.com
Vadym here - built the skills directory and marketplace infrastructure.
The core design decision was full composability. You can install multiple skills and they work together without conflicts. Each one adds a layer of expertise without overriding the others.
Seeing the first external skill builders package their knowledge into this system was the moment it clicked. The architecture works.
getviktor.com
Antoni here. I do growth at Viktor.
The honest version of why we built this: users kept telling us Viktor was great at pulling data but the analysis felt like it came from someone who read a blog post about Google Ads. Which, technically, it had.
With a skill installed it's a different conversation. Same data, same tools, but Viktor actually reasons like someone who's managed real accounts. The "oh wait, it caught that?" moment is real.
We started with ads because you can't fake results. ROAS improves or it doesn't. But this works for any domain where expertise compounds. If you're the person people DM when something breaks, your playbook probably belongs in the marketplace.