AI? Yes. 100%

Did we use AI to build Tape? Yes. 100%.

Look, we’ve seen the internet. Right now, "AI-generated programming language" is basically a synonym for low-effort slop—hastily prompted syntax with no underlying architecture, built by someone who learned to code yesterday just to farm stars on GitHub and karma on Reddit. If you're skeptical of Tape because of how it was built, we don't blame you. Skepticism is healthy. When we say 100% AI-assisted, we mean it’s deeply blended: at this point, the workflow is so integrated that we can't easily separate what was hand-typed from what was generated. So we own it.

Tape didn't start in a vacuum; it evolved from a game development effort streamed live on Twitch nearly 10 years ago. It began as an in-game stack-based scripting language for procedural assets, models, and textures. Over time, it grew into a tool to extend the game, and eventually, the language used to build the game itself. Drawing architectural inspiration from Robert Nystrom's Crafting Interpreters, Tape was born. Now it's been used inside Puida for developing internal tools, cross-platform CI runners, user interface, testing frameworks and as a hobby we experimenting with building a kernel with it.

Tape wasn’t dreamed up by an LLM. It’s built by Puida Oy, led by a developer with decades of professional software engineering experience—including work at Rovio, Aibidia, and KSML. That background includes designing internal programming languages across three different companies and building domain-specific languages for development long before LLMs ever existed.

We didn’t just type "write a compiler" and paste the output. We used AI intensively as a high-velocity force multiplier—generating boilerplate, writing test suites, exploring syntax iterations, and drafting repetitive parsing logic at breakneck speed. We use AI across all our products, logos, and designs. An LLM can write a parser, but it doesn't understand the core problem we are trying to solve: making games and writing software is hard. The vision, the constraints, and the architectural responsibility are 100% human-driven.

AI gave us the velocity to build Tape in a fraction of the time it would normally take. But the design, the intent, and the execution belong to a veteran engineer who actually understands how compilers work under the hood.

Don't judge Tape by the tooling we used to build it. Judge it by the compiler output, the architecture, and whether it solves a real problem. If it breaks or falls short, that's entirely on us.

Pauli Suuraho
Founder and CEO of Puida Oy, Creator of Tape