Exclusive: Roblox Overhauls AI Development Tools With New Agentic Capabilities to Streamline Game Creation

In an exclusive reveal to TechCrunch, Roblox has announced the launch of new agentic AI features built to help developers plan, build, and test games across its platform, headlined by a full revamp of Roblox Assistant — the platform’s natural-language AI tool for game development — to support creators through every step of the development workflow.

Roblox notes that traditional single-step AI tools, which generate a full output immediately from a user prompt, often fail to fully capture a creator’s original creative intent. To fix this gap, the company has rolled out an upgraded Planning Mode that reimagines Assistant as a collaborative development partner. This new mode can analyze a game project’s existing code and data model, ask targeted clarifying questions to fill in gaps, and convert raw creator prompts into fully editable step-by-step development plans.

Planning Mode guides developers through the full lifecycle of a project: drafting an initial outline for the game, collecting creator feedback to refine fine details, finalizing the development approach, and then executing the agreed-upon plan. Creators can tweak the plan and add extra context at this stage to ensure their vision is accurately reflected before any permanent changes are made to the project.

For example, if a creator prompts Assistant to “build a coin-collecting park minigame with a fountain and foliage,” the tool will not jump straight to generating assets. Instead, it may ask what visual style the creator wants for the park, offering options like cartoony, realistic, and fantasy. It may also ask how core assets (the fountain and foliage) should be created, giving creators choices to build from scratch, use pre-made models from the Roblox Creator Store, or combine both approaches.

Once a final plan is locked in, Planning Mode leverages Roblox’s full suite of AI tools to build out the game, including two new tools launched alongside the updated Assistant: Mesh Generation and Procedural Model Generation, both designed to cut down overall development timelines.

Mesh Generation simplifies adding fully textured meshes (3D objects) directly into the game world. Historically, developers rely on low-quality placeholder assets during early development to test how players will interact with the game space. With Mesh Generation, creators can quickly generate usable 3D models on demand, eliminating the need for low-quality temporary placeholders. For instance, creators can prompt Assistant to generate a campfire, add dynamic lighting to make it more realistic, and adjust the scene to be set at night all in a few simple steps.

Roblox also plans to launch Procedural Models soon, a tool that lets developers build editable 3D models via code and natural-language prompts through Assistant. Since Assistant understands 3D space and physical object relationships, creators can use prompts to place and resize assets relative to other objects already in the scene. Attributes like the number of shelves on a bookcase or the height of a staircase can be adjusted dynamically, creating reusable, editable building blocks that creators can refine and repurpose across projects.

“The launch of our agentic features in Roblox Studio reduces barriers between creative vision and execution,” Nick Tornow, Roblox’s Senior Vice President of Engineering, said in a statement to TechCrunch. “Creating with Planning Mode and our Procedural Generation tools is a powerful new method for creators to turn their concepts into gameplay. Assistant works as a multi-step, collaborative development partner — accelerating the process of planning, building, and testing, so creators can get from idea to reality faster.”

As Planning Mode executes the development plan, it uses built-in playtesting tools to automatically audit the build: it reads output logs, captures in-game screenshots, uses keyboard and mouse inputs to test design and gameplay, flags bugs, and feeds that data back to Assistant so issues can be fixed automatically.

“With the new capabilities across planning, building, and testing, 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,” Roblox explained in a public blog post.

Looking ahead, Roblox confirmed it is developing new upgrades including support for multiple AI agents working in parallel on the same project, the ability to run long, complex development workflows in the cloud, and automated handling for advanced tasks like full-code writing, end-to-end testing, and creating more realistic in-game characters. The company is also working to enable seamless integration between Roblox Studio and popular third-party AI tools including Claude, Cursor, and OpenAI Codex, to give creators more flexibility to use the tools they already prefer.