Playdate's 1-Bit 3D: A Dev Defies Limits With a Custom Renderer
Cristina Ramos built a 3D renderer from scratch for Panic’s Playdate — a monochrome handheld with a 4 MHz ARM Cortex-M7 and 16 MB of RAM — because multiple people told her it could not be done. The result is Katavatis, a first-person “Metroidbrainia” that renders indoor spaces and outdoor cityscapes in crisp 1-bit black and white, with the same engine already confirmed running DOOM. For a device that most assumed would never leave 2D sprite land, this is a technical challenge thrown down and won.
Why a Software Renderer on Cortex-M7 Changes the Playdate’s Ceiling
Ramos’s renderer sits in a visual sweet spot between the 3DO and Sega Saturn, despite having no dedicated 3D hardware. She solved polygon rasterization and perspective projection in software, confirming the 4 MHz chip can push textured polygons when every cycle is accounted for. The Playdate’s battery life (up to 8 hours for most games) holds up here: the small monochrome framebuffer keeps memory bandwidth low even when the CPU is pinned on render work. No battery anxiety for on-the-go play.
Katavatis: Deep-Sea Metroidbrainia with a PC Release
The game is set in an abandoned deep-sea facility with a climate-apocalyptic story, using knowledge-based Metroidvania progression — the player unlocks understanding, not gear. Ramos has published a devblog on the renderer’s creation and says the engine is now “good enough” for gameplay prototypes. Katavatis will release on Playdate, Steam, and Itch.io, letting PC players experience the 1-bit 3D aesthetic at higher resolutions. No release date yet, but code snippets are already live on YouTube. For Playdate owners, Katavatis is the most technically ambitious title on the platform. For the broader handheld community, it proves that ultra-low-spec hardware — the kind usually written off for 2D-only work — can still deliver spatial depth and first-person immersion, much like the Pepsiman browser recompilation and the Game Bub FPGA handheld demonstrated the untapped potential of existing hardware. The ceiling is not the silicon; it is whether developers are willing to write their own rasterizer.