Configuring Claude Code for Privacy and Noise Control

If you're using Claude Code for development work, you might want to minimize telemetry, disable promotional messages, and keep your Git history clean. Here's how to configure Claude Code for maximum privacy and minimal noise.

Telemetry and Git Spam

Out of the box, Claude Code includes:

• Feedback surveys that pop up mid-session
• Analytics and error reporting
• Git commit footers with promotional links
• Co-author attributions in commits
• Extra "flavor text" from non-essential LLM calls

For privacy-conscious developers or those who prefer a quieter workflow, these features can be distracting or unwanted.

Privacy and Noise Control

Create or edit your Claude Code settings file. The location varies by OS:

• macOS/Linux: ~/.config/claude-code/settings.json
• Windows: %APPDATA%\claude-code\settings.json

Here's a fully-configured setup:

{
  "$schema": "https://json.schemastore.org/claude-code-settings.json",

  "alwaysThinkingEnabled": true,

  "feedbackSurveyState": {
    "lastShownTime": 1754109357477
  },

  "includeCoAuthoredBy": false,

  "env": {
    "CLAUDE_CODE_DISABLE_FEEDBACK_SURVEY": "1",
    "DISABLE_TELEMETRY": "1",
    "DISABLE_ERROR_REPORTING": "1",
    "DISABLE_NON_ESSENTIAL_MODEL_CALLS": "1"
  }
}

What Each Setting Does

Git Cleanup

"includeCoAuthoredBy": false

Prevents Claude Code from adding these lines to your commits:

🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>

Note: While this is the official setting to keep your commit history clean, it does not always work. Claude will still sometimes add those lines to your commits. See this GitHub Issue.

Privacy Controls

Environment variables in the env block control data collection:

• CLAUDE_CODE_DISABLE_FEEDBACK_SURVEY — Stops the "How is Claude doing?" survey from interrupting your workflow
• DISABLE_TELEMETRY — Opts out of analytics and usage tracking
• DISABLE_ERROR_REPORTING — Prevents automatic error reports (like Sentry)

Noise Reduction

DISABLE_NON_ESSENTIAL_MODEL_CALLS — Reduces unnecessary LLM calls for tips, banter, and flavor text. Claude gets straight to work without the small talk.

Optional: Keep Extended Reasoning

"alwaysThinkingEnabled": true — This has nothing to do with privacy. It enables Claude's extended reasoning mode, which can improve code quality. Safe to keep or remove based on preference.

Account-Level Privacy

The settings above control the CLI tool, but there's one more step for complete privacy:

1. Open claude.ai
2. Go to Settings → Privacy
3. Toggle "Help improve Claude" to OFF

This prevents your conversations and code from being used for model training.

Benefits

With this configuration, you get:

• ✅ No telemetry or tracking
• ✅ No interrupting surveys
• ✅ Clean Git history
• ✅ Quieter, more focused workflow
• ✅ Full control over your data

Notes

• These settings don't disable auto-updates, you'll still get the latest Claude Code features and fixes
• The feedbackSurveyState timestamp is a legacy leftover and harmless to keep

What I'm building

Delegate tasks. Get software.

Give Vroni a GitHub issue, bug report, spec, or rough idea. It reads the repo, plans the change, writes code, runs checks, and works toward a review-ready pull request.

Take a look at vroni.com