Compiler Explorer

Compiler Explorer Data Privacy and Usage Policy

• Core Functionality: Compiler Explorer is a non-commercial, volunteer-run project (led by Matt Godbolt) that compiles source code in temporary environments. Code is deleted within minutes of compilation.
• AI Integration (Claude Explain):
  • Users must provide explicit consent before code is sent to Anthropic for explanation.
  • Anthropic does not use this data for model training.
  • The "no-ai" directive in source code prevents AI processing.
• Data Retention & Security:
  • Debugging: With user consent, code may be stored for up to one week if a system error occurs.
  • Caching: Results are cached via secure hashes; original source code cannot be reconstructed from these hashes.
  • Shared Links: "Full" links store code in the URL; "Short" links store state on Compiler Explorer servers, accessible only by administrators.
  • Logs: Web logs use semi-anonymized IP addresses; Amazon infrastructure logs (full IPs) are deleted after 32 days; Sentry error reports are kept for 90 days.
• Analytics: Usage patterns are tracked via code fingerprints (hashes) and kept for up to one year to improve the service. Aggregate statistics may be shared publicly.
• Execution: Code execution occurs in a locked-down, isolated environment designed to prevent information leakage.
• User Control:
  • Settings, source code, and UI configurations are stored locally in the user's browser.
  • Users can request deletion of shared links or data by emailing privacy@compiler-explorer.com.
  • The project is open source; users can host their own private instances via GitHub.
• Compliance: The service adheres to GDPR by avoiding long-term storage of personal information and utilizing IP anonymization.