xAI Just Dropped Their Claude Code Killer

xAI Just Dropped Their Claude Code Killer

Introduction to Grok Build

Overview of Grok Build

  • XAI has launched Grok Build, a new tool in beta that aims to simplify coding by allowing users to create a fully playable retro arcade game from a single prompt.
  • The command-line interface (CLI) is designed for speed and efficiency, enabling users to initiate it with just one command.
  • Grok Build integrates seamlessly with existing tools and files, supporting various formats like agents.md files and plugins.

Features of Grok Build

  • Upcoming changes in Claude's API credits may impact user experience; community feedback suggests mixed feelings about these changes.
  • The goal is to provide an unrestricted coding experience through the CLI, which Grok Build appears poised to deliver.

Setting Up Grok Build

Installation Process

  • The installation process involves copying a one-line installer command into the terminal, which successfully sets up the XAI CLI.
  • User authentication is required via a browser window that opens during setup.

Initial Configuration

  • Users can create directories for their projects directly within the terminal environment, enhancing workflow efficiency.

First Project: Building an Arcade Game

Prompting for Game Creation

  • The initial prompt requests the creation of an Asteroids-style retro arcade game playable in browsers without frameworks or external libraries.
  • Users can choose different modes such as "Grok plan mode" for better control over project development.

Approval Process

  • After generating a plan for the game, users have options to approve or revise it before proceeding with code generation.

Game Development Experience

Code Generation and Gameplay Testing

  • Within five minutes, Grok Build generates a playable version of the requested arcade game featuring basic controls and gameplay mechanics.
  • The tool demonstrates high capability by producing quality output quickly compared to traditional coding methods.

Enhancements Through Sub-agents

Parallel Task Management

  • A follow-up prompt requests six major upgrades for the game, delegating tasks to parallel sub-agents for efficient processing.
  • Multiple sub-agents are created simultaneously, each handling specific enhancements like enemy designs and power-ups.

Results of Parallel Processing

  • Each sub-agent completes its assigned task independently while maintaining high-quality outputs without conflicts among them.

Advanced Features and Commands

Utilizing Programmatic Commands

  • Users can execute commands programmatically using grok-p, allowing further automation in managing project files like README documentation.

Real-time Feedback Mechanism

  • While tasks are being processed by sub-agents, real-time responses are provided on other queries related to sound frequencies used in games.

Conclusion: Evaluating Grok Build's Performance

Final Thoughts on Game Development

  • With minimal input and effective use of sub-agents, users can create complex games complete with graphics and soundtracks efficiently.
Video description

πŸš€ Try Grok Build: https://mrc.fm/xcli xAI just dropped Grok Build, their answer to Claude Code, and I had to put it to the test so I gave it one prompt and watched it vibe code a fully playable retro arcade game from an empty folder in about five minutes. Then I pushed it further with six parallel subagents building UFOs, power ups, chiptune music, screen effects, a leaderboard and an achievement system all at the same time, ran headless grok -p commands from a second terminal to generate a README on the fly, and tested the polish skill that xAI's own engineers use. The slick Rust built TUI, mouse support, Plan Mode and the fact that it reads your existing CLAUDE.md and MCP setup with zero configuration might genuinely change how I code and at $99 for six months it's half the price of my Claude Max plan. Let me know in the comments what you'd build with it. 0:00 Intro & Grok Build Beta 0:18 Features & Overview (Claude Code Comparison) 1:35 Upgrade, Installation & Terminal Setup 3:04 First Prompt, Plan Mode & Generating the Game 5:24 Testing the Base Game & Context Limits 6:22 Upgrading with 6 Parallel Sub-Agents 7:46 Testing the Upgraded Gameplay 8:10 Grok -p Programmatic Commands (Readme Generation) 9:30 The /btw (By The Way) Command 10:12 Final Gameplay, Polish & Outro