SpaceXAI has open-sourced Grok Build, its AI coding assistant powered by Grok 4.5, following revelations that the tool was transmitting user code and authentication credentials to company servers without explicit consent. The move marks a significant transparency shift for Elon Musk's artificial intelligence company and comes as the executive also announced plans to open-source the entire codebase of X (formerly Twitter) after a security review.

What Happened With Grok Build's Data Transmission?

Grok Build, released in May 2026, was initially designed to help developers write code faster by automating routine coding tasks. However, the tool had a critical default setting that caught users off guard. When the beta version launched, the data upload feature was enabled by default, meaning user codebases and authentication credentials were being sent to SpaceXAI's servers without users explicitly opting in.

The company discovered this issue and took action on July 12, 2026, by disabling the data upload feature by default and deleting all previously uploaded coding data. SpaceXAI acknowledged the problem in a statement, noting that while the feature was enabled by default initially, the company maintained "Zero Data Retention" policies when users disabled uploads through the command-line interface.

Why Did SpaceXAI Open-Source the Code?

Rather than simply fixing the issue quietly, SpaceXAI took the more radical step of releasing Grok Build's source code publicly under the Apache License 2.0 on July 15, 2026. The code is now available on GitHub in the "xai-org/grok-build" repository, allowing developers to inspect exactly how the tool works and compile it in their own environments.

This transparency move serves multiple purposes. By publishing the source code, developers can now understand the complete operational principles of how Grok Build functions, from context building to tool distribution. For developers using advanced features like Agent Skills, plugins, hooks, and MCP servers, the published code acts as a "trusted reference" for understanding how each component loads and operates.

What Can Developers Do Now With Grok Build?

• Local-First Operation: Developers can now run Grok Build entirely on their own machines when paired with local inference infrastructure, eliminating the need to send any data to SpaceXAI's servers.
• Code Inspection: The open-source release allows developers to review the complete source code and verify that the tool operates as advertised, addressing trust concerns from the data transmission incident.
• Custom Extensions: Developers can extend Grok Build's functionality through Skills, MCP servers, and sub-agents, with full visibility into how these components integrate with the core system.
• Security Auditing: SpaceXAI is encouraging security researchers to report vulnerabilities through HackerOne, creating a formal channel for identifying and fixing potential issues.

How Does This Reshape the AI Coding Assistant Market?

The open-sourcing of Grok Build could significantly disrupt the competitive landscape dominated by tools like GitHub Copilot and Cursor. A fully local, source-available alternative may appeal strongly to developers who prioritize security and privacy over convenience. Unlike proprietary competitors, developers using Grok Build can now verify that their code never leaves their machines unless they explicitly choose to upload it.

This move also reflects a broader trend toward transparency in AI development. By allowing independent review of the code, SpaceXAI is attempting to rebuild trust after the data transmission controversy. The company emphasized that "by reading the source code, you can understand the complete operational principles," positioning transparency as a competitive advantage.

What About Elon Musk's Plan to Open-Source X?

In parallel with the Grok Build release, Musk announced on July 15, 2026, that X's entire codebase will be open-sourced once a security vulnerability review is complete. This goes far beyond X's previous decision to publish only the core recommendation system powering its "For You" feed.

The plan includes bringing in independent third-party reviewers to verify that systems running in production actually match the published source code. This addresses a fundamental challenge with open-sourcing: external parties cannot normally confirm that the public code is identical to what is actually running on company servers. However, Musk has not disclosed the release date, repository location, software license, or technical scope of what will be included.

The scope could be substantial. X's product portfolio now extends beyond its social feed to include XChat, which features encrypted messaging, calling, and Grok integration. All of these systems could potentially fall within the disclosure scope, raising both transparency and security considerations.

What Are the Risks of This Open-Source Strategy?

While open-sourcing code increases transparency, it also carries real risks. Publishing source code for systems with significant privacy and security implications could expose credentials, private data, internal security controls, and infrastructure details. SpaceXAI and X will need to carefully review their code before publication to remove sensitive information.

Additionally, the timeline for X's full open-sourcing remains vague. Musk has not specified when the security review will be complete or how the company will handle ongoing updates after publication. These unanswered questions leave developers uncertain about when they can expect access to the full codebase.

The Grok Build open-sourcing represents a significant bet on transparency as a trust-building mechanism. Whether this strategy succeeds in winning over privacy-conscious developers will likely influence how other AI companies approach similar decisions in the coming months.