[[["わかりやすい","easyToUnderstand","thumb-up"],["問題の解決に役立った","solvedMyProblem","thumb-up"],["その他","otherUp","thumb-up"]],[["必要な情報がない","missingTheInformationINeed","thumb-down"],["複雑すぎる / 手順が多すぎる","tooComplicatedTooManySteps","thumb-down"],["最新ではない","outOfDate","thumb-down"],["翻訳に関する問題","translationIssue","thumb-down"],["サンプル / コードに問題がある","samplesCodeIssue","thumb-down"],["その他","otherDown","thumb-down"]],[],[],[],null,["[](/showcase) \nShare\nMay 16, 2025 \n\nCartwheel Advances Character Animation Using the Gemini API \nJonathan Jarvis\n\nCo-founder \\& CTO \nAndrew Carr\n\nCo-founder \\& Chief Scientist \nVishal Dharmadhikari\n\nAI DevRel \n\nCartwheel Advances Character Animation Using the Gemini API \n\nCartwheel stands at the forefront of next-generation animation, offering a platform where natural language inputs (e.g., \"jump,\" \"salsa dance spin\") direct character actions. This innovation improves how creators animate for video, gaming, advertising, and social media. Cartwheel integrates Imagen 3 for intuitive text-to-character design, and harnesses Gemini 2.5 Pro Preview to develop sophisticated animation-specific algorithms, explore complex codebases, and accelerate platform optimization, empowering users to expand their creative storytelling.\n\nOvercoming R\\&D Hurdles in Generative 3D Motion \n\nDeveloping a cutting-edge animation platform requires not only user-friendly design tools but also continuous innovation in backend algorithms and efficient codebase management. Cartwheel faced the challenge of rapidly prototyping and implementing complex animation features, optimizing their extensive codebase for performance, and enabling developers to quickly understand and contribute to a large-scale system. They needed a powerful AI partner to supercharge their development efforts.\n\nHow Cartwheel Uses Imagen 3 and Gemini 2.5 Pro Preview \n\nCartwheel uses the Gemini API to improve its user-facing creation pipeline and to assist its internal development processes.\n\n\n\u003cbr /\u003e\n\n\n- **AI-Driven Character Conception with Imagen 3:** Cartwheel integrates Imagen 3 text-to-image capabilities, allowing creators to use text prompts to conceptualise and generate unique character visuals directly within the platform. These custom-designed characters can be animated using natural language, then exported fully rigged and ready for industry-standard 3D editing software like Maya or Blender, enabling integration into professional production environments.\n- **Assisting R\\&D and Optimization with Gemini 2.5 Pro Preview:** The Cartwheel team uses Gemini 2.5 Pro Preview internally to support its engineering and research workflows:\n - **Developing Algorithms:** Gemini 2.5 Pro Preview is employed to help develop animation-specific features, including algorithms for motion blur for data synthesis, Laplacian pyramids for blending, and auto IK (Inverse Kinematics) for in-editor posing.\n - **Code Exploration \\& Debugging:** The team uses Gemini 2.5 Pro Preview to explore complex codebases, generate ideas for new experiments, and assist in efficiently debugging edge cases, such as intricate camera rotation errors.\n - **Leveraging Long Context for Codebase Operations:** Gemini 2.5 Pro Preview's long context capabilities are used to operate over Cartwheel's entire codebase. This enables developers to, for example, bring new codebases (like the front-end) into context to help add features or ask high-level questions about the system's architecture and functionality.\n\n\u003cbr /\u003e\n\nEmpowering Animation Workflows \n\nBy using Gemini 2.5 Pro Preview to assist with algorithm development, code exploration, and debugging, Cartwheel aims to enhance its development processes and accelerate platform optimisation. The integration of Imagen 3 provides users with a streamlined way to generate character visuals from text, while Gemini 2.5 Pro Preview serves as a tool to aid the developers building the underlying technology. These integrations support Cartwheel's goals of making animation workflows faster and more accessible, enabling uses like:\n\n\n\u003cbr /\u003e\n\n\n- Accelerating asset creation for games and video, allowing teams to focus on core design.\n- Facilitating efficient marketing and social media content generation.\n\n\u003cbr /\u003e\n\n\u003e \"Animation is one of the purest ways to get a story out of your head and show it to the world. We're building tools to make that easier and more accessible for everyone!\"\n\n- Jonathan Jarvis, CEO / Co-founder, Cartwheel\n\nBuilding the Future of Animation \n\nCartwheel's integration of Imagen 3 and Gemini 2.5 Pro Preview highlights how Google AI can be applied to enable new user-facing features and assist the complex development work required to build next-generation tools.\n\n\n\u003cbr /\u003e\n\n\nExplore the [Gemini API documentation](https://cloud.google.com/gemini/docs/overview), learn about [Imagen 3](https://ai.google.dev/gemini-api/docs/image-generation#imagen), and get started in [Google AI Studio](https://ai.dev\").\n\n\n\u003cbr /\u003e\n\n\nCartwheel is a participant in [Google's AI Futures Fund](https://labs.google/aifuturesfund) that invests in and collaborates with ambitious startups building what's next in AI. \n\nRelated case studies \n[Vela Partners\nVela Partners uses Grounding with Google Search for Deeper, Faster Insights](/showcase/vela) [Jolt AI\nJolt AI uses Gemini API for fast, accurate understanding of large codebases.](/showcase/joltai) [Optimal AI\nOptimal AI Uses the Gemini API to Cut Code Review Times by 50%](/showcase/optimalai)"]]