Thirty Days In: What We Built Before We Shipped

Thirty days ago, I gave myself a simple and slightly unreasonable goal: release a game.

I did not get Wordulate into the App Store by June 30, 2026. That is the plain version. The more useful version is that the last 30 days turned an idea, a repo, and a lot of conviction into a real independent game studio with a working mobile prototype, a growing production system, and a much clearer understanding of what it will take to ship well.

I am proud of that. Not in the inflated, victory-lap way. More in the tired, grounded, "this is real now" way.

Snarly Owl Games LLC exists. The domains are claimed. The public handles are mostly in place. The first game has a name, a shape, a prototype, and a backlog that looks less like a fantasy and more like a production plan. Wordulate is not released yet, but it is no longer imaginary.

And the strange, important part is this: the whole thing has been built primarily with Codex and ChatGPT as part of the working team.

Why I Started

Snarly Owl Games is an independent game studio based in West Virginia. The short version of the studio description is simple: games for brains.

The longer version is more personal.

I wanted to build games that respect thoughtfulness. Games that care about skill without becoming joyless. Games that can be played casually but still reward attention, improvement, and the satisfying little click of seeing something you did not see a minute ago.

Wordulate is the first attempt at that. It is a mobile word puzzle built around the feeling that word games can do more than count attendance. Streaks are powerful. Points are useful. But I keep coming back to a different question: how do we help a player understand the quality of a solve? How do we make a word game feel smart, fair, and worth returning to without turning it into homework?

That question became the product center. Around it, the company started forming.

In the same month that I was working on the game, I was also doing the ordinary company-formation work: choosing the final legal name, filing the West Virginia LLC paperwork, getting an EIN, buying domains, claiming social handles, writing the first studio bio, and starting the early business checklist that every founder eventually has to stop avoiding.

I also talked with Mary at the West Virginia small-business support office about grant opportunities. That conversation mattered because it made the company feel less like a private dare and more like something that belongs in an actual local business ecosystem. There is still plenty of paperwork left, but the studio has crossed the line from "one day" into "open for work."

The 30-Day Reality

The daily rhythm was not glamorous.

Most days started with 4 to 6 hours on the company and the game: product decisions, code, docs, build issues, testing, store-account setup, strategy, names, systems, and the thousand tiny judgments that make a prototype less vague. Then the day shifted into another 4 to 6 hours of renovating the house in Maryland that we are selling, packing up the old life, and moving toward a smaller one in West Virginia.

That context matters to me because it is part of the story. This was not a 30-day sprint in a clean studio with a perfect calendar. It was building a company in the morning and sanding, hauling, fixing, painting, or moving in the afternoon. It was trying to make something new while actively taking apart the old shape of our life.

The original goal was an app-store release. The real outcome was a foundation.

Some of the blockers were mundane and very real. The app-store path ran into identity and account setup friction, including being blocked on official West Virginia license timing. That was frustrating, but also clarifying. Store release is not just a button at the end of development. It is its own production lane, with identity checks, business records, developer accounts, build signing, store metadata, policies, and review readiness.

Trying to ship exposed the work. That is useful, even when it is annoying.

What We Built

Wordulate now has a playable mobile prototype. The current build is an Expo/React Native app in a TypeScript monorepo, with shared game logic, product docs, architecture docs, backend packages, and a growing set of validation tools.

The first prototype moved through several important stages:

• We turned the idea into a structured product direction instead of a loose concept.
• We built the first playable mobile session so the game could be judged by touch, timing, feedback, and pacing.
• We tightened responsiveness, board feel, audio, visual identity, and HUD clarity.
• We created an Android preview build path so the prototype can start traveling beyond the development machine.
• We added backend foundations for server-backed gameplay, replay, telemetry, and Daily Puzzle work.
• We started formalizing the word pipeline, validation checks, and playtest readiness gates.

That list is dry on purpose. It is the kind of progress that does not always photograph well, but it is the difference between a demo and a production effort.

One of the biggest lessons of the month was that "playable" is not the same as "ready to learn from." A prototype can technically run and still produce bad signals because the feedback is vague, the build path is fragile, the word list is under-tested, or the first session asks too much forgiveness from the player. A lot of the last 30 days was about making the game sturdy enough that future feedback will mean something.

That is why we invested in boring-sounding things: smoke tests, level validation, backend contracts, telemetry events, Android build lanes, replay behavior, task checklists, and docs that keep the work from disappearing into chat memory.

The work is not finished. But it is much more inspectable than it was.

What It Cost

The part of this month that I think more people should see is the cost.

So far, the direct cash cost to build Snarly Owl Games and get Wordulate to this stage is about $400 for the first month.

• $180 for the LLC registration
• $50 for the two business licenses for ChatGPT and Codex
• $40 for 12 months of website hosting
• $20 for a monthly Figma license after the free plan ran out of MCP credits
• $70 for internet access
• $40 for domain registration for the year

Not all of that is monthly operating expense. The LLC registration is a formation cost, while hosting and domains are annual costs paid up front. The rough recurring monthly operating cost right now is closer to $140: ChatGPT and Codex, Figma, and internet access.

We have also used the free versions of Scenario and ElevenLabs for early art and sound exploration. The art has been created with ChatGPT, with some pieces, including the logo, needing more direct human touch in Figma.

That cash number is not the full economic cost of making a game. If I count my own time at a livable West Virginia wage of $30 per hour, and estimate 4 hours per day for 30 days, that is 120 hours of work, or about $3,600 in founder labor.

Put another way, the first month cost about $400 in actual cash out of pocket and about $3,600 in unpaid founder wages. If the company had already been making enough money to pay me for that time, the all-in first-month cost would be about $4,000 before counting existing hardware, future store fees, taxes, legal help, contractors, marketing, or the opportunity cost of choosing to build this instead of doing something safer.

But the number still matters. A playable mobile game prototype, a real company, a public website, a production workflow, and the first publishing lane were not built with a giant budget. They were built with a small amount of cash, a lot of judgment, and tools that made one person more capable across product, code, art direction, operations, and QA.

That is part of the founder vision for Snarly Owl. Game development in the United States has to get its costs under control if we want to stay competitive globally. Not by making worse games, exploiting people, or pretending craft does not matter. By building smaller, more focused studios that can move carefully, spend responsibly, and use modern tools to stay in the fight.

I am increasingly convinced that the future of games includes more small studios, not only larger ones. AI tools are not the whole answer, but they are a real unlock. They can lower the cost of exploration, reduce the gap between idea and prototype, and help a tiny team compete in places where the old model would have required money we did not have.

The AI Part

I want to be direct about this because I think it is one of the most interesting parts of the project.

Wordulate has been built primarily with Codex and ChatGPT.

Not as a novelty. Not as a "look, AI made a game" trick. More like a new small-studio operating model.

ChatGPT has been the thinking room: product critique, founder reflection, business planning, naming, positioning, design stress tests, and the place where messy ideas get sharpened enough to become work.

Codex has been the production partner: reading the repo, editing code, writing docs, creating scripts, running tests, opening Android paths, keeping task lists updated, and turning decisions into actual files. We have used Codex to run automated smoke tests, inspect failures, validate backend behavior, and watch whether the game still launches under the conditions a tester would actually encounter.

That does not remove human judgment. If anything, it makes judgment more important. The AI can produce a lot of motion quickly, but the founder still has to decide what matters, what is safe to reveal, what feels wrong, what should be cut, and what kind of game this is trying to become.

The best version of this workflow has not felt like outsourcing the soul of the project. It has felt like increasing the surface area of a very small studio. A single founder can now keep product strategy, mobile implementation, backend planning, QA lanes, publishing notes, and business operations in motion in a way that would have been difficult to imagine not long ago.

There are tradeoffs. AI can move faster than your ability to review if you let it. It can create too much documentation. It can make the project look more complete than it is if you confuse artifacts with readiness. This month taught me to treat Codex and ChatGPT as powerful collaborators, not magic authority.

The job is still to ship a good game.

Where Wordulate Is Now

As of July 1, 2026, Wordulate is in prototype-to-playtest development.

The current focus is not making a louder promise. It is making the first experience clearer, more responsive, more testable, and easier to trust. The Android playtest path is closer than the iOS store path because Android distribution is less blocked by the account issues I hit this month. Daily Puzzle work is active because a word game needs a return ritual that feels native to the product, not bolted on later. Backend and telemetry work are active because friendly playtests are only useful if we can understand what happened.

There is still a lot to do:

• tune the early levels
• improve clue and booster presentation
• complete the Daily Puzzle result, reward, and share flow
• keep hardening the Android preview build
• unblock the Apple developer path
• collect real tester feedback
• make the first session feel less like a promising prototype and more like a game people want to keep playing

That is not the finish line I wanted after 30 days. It is, however, a real map.

What Comes Next

The next phase is simple to describe and difficult to do: get Wordulate into more hands, learn honestly, and keep reducing the distance between the game we can feel in our heads and the game a player can feel in their hands.

I also want to document the process more publicly from here. Not just the wins. The blockers too. The LLC paperwork. The store-account friction. The conversations about grants. The way AI changes the practical limits of a small studio. The tension between building a game and rebuilding a life around it.

I missed the first deadline. I did not miss the point.

Thirty days in, Snarly Owl Games is real. Wordulate is playable. The workflow is unusual enough to be worth studying. The company has roots in West Virginia now. And the next goal is no longer to prove that I can start.

It is to keep going until the game is ready.