sebattfg/ZeroScript-Free

ZeroScript Free: Turn ChatGPT, DeepSeek, Gemini, Kimi, GLM, Qwen, Arena or Meta AI into a Roblox Studio agent. Browser extension + local bridge. Read/edit scripts, run Luau, generate assets, all from chat. Robust agentic loop. No terminal needed.

What it solves

ZeroScript provides a way to control Roblox Studio using popular AI chat interfaces (like ChatGPT, DeepSeek, and Gemini) without needing an API key, a terminal, or coding knowledge. It transforms a standard browser-based AI chat into an active agent that can manipulate the game engine directly.

How it works

The system uses a three-part architecture: a browser extension that integrates into the AI chat page, a local "Bridge" application running on the user's PC, and the Roblox Studio MCP server. When a user types a request in the AI chat, the extension sends commands to the local Bridge, which then drives Roblox Studio via its built-in Model Context Protocol (MCP) server.

Who it’s for

Game developers using Roblox Studio who want to automate script writing, asset generation, and game tree inspection using AI agents.

Highlights

  • Multi-LLM Support: Works with DeepSeek, ChatGPT, Gemini, Kimi, GLM, Qwen, Arena, and Meta AI.
  • Studio Integration: Can read/edit scripts, run Luau code, inspect game instances, and control play-testing.
  • Asset Creation: Capable of generating meshes, materials, and models, and browsing the Creator Store.
  • No-Code Setup: Operates via a browser extension and a simple bridge executable, removing the need for API keys.
  • Persistent Memory: Remembers project details across different sessions within a specific place.

Related

  • Project
    osolmaz/pi-workflows

    A workflow extension for the Pi coding agent that lets users define multi-step agent tasks as TypeScript graphs with durable execution and live terminal visualization.

  • Project
    corespeed-io/zypher-agent

    Zypher Agent is a framework for building production-ready AI agents that use reactive reasoning loops to dynamically execute tasks within applications.

  • Project
    Jeidoban/Ironsmith

    A macOS menu bar app that uses AI to generate, build, and export native SwiftUI applications from simple text descriptions.

  • Project
    Puiching-Memory/TAAC_2026

    An experimental workspace for large-scale recommendation systems that unifies sequence modeling and feature interaction for conversion rate prediction.