
Create your own visual novel using AI for story and image generation
What tools did you use to create your project?
How much experience does your group have? Does the project use anything (art, music, starter kits) you didn't create?
Our group has mixed experience levels. A few members have prior experience with Python programming and AI APIs, while others are new to both game development and Ren’Py. This project was our first time combining visual novel scripting with generative AI tools. We learned to integrate APIs (Gemini for story generation and Stability AI for image creation), structure a Ren’Py project, and manage code collaboration across multiple scripts. Overall, it was a learning-focused project where everyone contributed to both coding and creative decisions.
What challenges did you encounter?
One of our biggest challenges was integrating multiple systems — connecting Ren’Py with external APIs like Gemini and Stability AI while keeping the game flow stable. Managing asynchronous image generation during the loading and gameplay phases required careful pipeline logic to ensure the next scenes were ready before the player reached them.
We also ran into issues with AI prompt consistency, such as maintaining character appearance across scenes and ensuring the generated story stayed coherent and simple. Another challenge was handling image resolution and scaling in Ren’Py, since different aspect ratios sometimes caused layout problems.
Finally, collaboration itself was a challenge — dividing tasks among six people, keeping the code organized, and managing API keys and dependencies across machines took planning and communication.