Roblox: AI Transforms Game Creation with New Tools

Roblox, the well-known online gaming platform, recently announced a series of significant improvements to its AI Assistant, a tool designed to support developers in game creation. These new features aim to transform the Assistant into a collaborative partner, capable of guiding creators through the various stages of development.

Roblox revealed these updates exclusively to TechCrunch, highlighting the transformation of its AI Assistant into a natural language game development tool. The goal is to support creators throughout the development process, from planning to building, and finally to testing the games.

Enhanced Planning Mode

One of the main updates is the enhanced "Planning Mode." Unlike traditional AI tools that produce one-step solutions, this mode allows the Assistant to collaborate with developers by analyzing a game's code and data model. It asks clarifying questions and translates requests into editable action plans, ensuring that the creator's intent is respected.

Roblox emphasizes that AI tools that simply take a request and produce a one-step solution often fail to truly capture a creator's original intent. This is why the Planning Mode was designed to help developers create a plan for their game, receive feedback to refine details, finalize the approach, and then implement that plan. Creators can adjust the plan and add context to ensure their intent is clearly reflected before any changes are made.

For example, if a developer wants to create a park mini-game with specific elements like a fountain and vegetation, the Assistant can ask for clarifications on the desired visual style, offering options such as cartoon, realistic, or fantastical. The Assistant can also inquire how the creator wants the park elements, like the fountain and vegetation, to be created, suggesting options like building from scratch, using models from the Creator Store, or a mix of both.

Mesh Generation and Procedural Models

Roblox has also introduced two new tools to accelerate development: Mesh Generation and Procedural Models. Mesh Generation allows for the rapid addition of fully textured 3D objects into the game, facilitating the creation of realistic models without relying on low-quality placeholder assets.

Roblox indicates that during the early stages of development, developers often create placeholder assets to understand how players will interact with the world. With Mesh Generation, creators can quickly create 3D models instead of having to rely on low-quality replacements. For instance, creators can ask the Assistant to generate a campfire, then add light to make it more realistic, and finally set the scene at night.

Procedural Models, on the other hand, enable developers to create modifiable 3D models with code. The Assistant, understanding 3D space and physical relationships, can dynamically adjust attributes such as the size of objects or the number of shelves in a library. Attributes like the number of shelves in a library or the height of a staircase can be dynamically adjusted, creating modifiable building blocks that can be refined and reused elsewhere.

An Optimized Development Process

Nick Tornow, Senior Vice President of Engineering at Roblox, emphasized that these new features reduce the barriers between creative vision and execution. The Planning Mode and Procedural Generation tools allow creators to transform their concepts into gameplay more quickly and efficiently.

“The launch of our agentic features in Roblox Studio reduces the barriers between creative vision and execution,” said Nick Tornow, Senior Vice President of Engineering, in a statement to TechCrunch. “Creating with the Planning Mode and our Procedural Generation tools is a powerful new method for creators to turn their concepts into gameplay. The Assistant acts as a collaborative development partner in multiple stages — speeding up the planning, building, and testing process so that creators can move from idea to reality more quickly.”

Additionally, the Assistant uses game testing tools to verify design and gameplay, identify bugs, and provide feedback for automatic corrections. Roblox is also working on integrating multiple AI agents capable of handling complex tasks in parallel, as well as compatibility with third-party tools like Claude, Cursor, and Codex.

As the Planning Mode executes according to the plan, it will utilize game testing tools to read output logs, capture screenshots, use inputs like a keyboard and mouse to verify design and gameplay, and identify bugs while providing feedback to the Assistant so it can correct them automatically.

“With the new capabilities across planning, building, and testing, the Assistant is better at using agentic loops to test different aspects of the game, surface suggested solutions, and then incorporate the results into future planning loops, creating a self-correcting system that becomes more accurate over time,” explained Roblox in a blog post.

Roblox also announced that it is working to enable multiple AI agents to work together in parallel, execute long and complex workflows in the cloud, and manage tasks such as coding, testing, and creating more realistic game characters. It also aims to ensure that creators can seamlessly use Claude, Cursor, Codex, and other third-party tools with Roblox Studio.