Montpellier Platformer built with Cursor, Codex, GPT-5.5

The interesting part here is less the game itself than the workflow: a short-jam game built end to end with an AI coding stack that now looks genuinely production-capable for lightweight creative projects.

• –Shows how fast AI coding tools can compress a small game project from idea to playable deployment
• –Cursor plus Codex plus GPT-5.5 is a strong signal that agentic coding is moving beyond code completion into full project assembly
• –Jam-style projects are a good proving ground because scope is constrained and iteration speed matters more than polish
• –The “Montpellier” theme gives the project identity, but the bigger story is the workflow, not the game genre
• –This is relevant to developers watching AI-assisted game dev, but it’s still a showcase project, not a commercial product launch