How do you manage feature flags in production without cluttering the codebase?

I've been reading about feature flags and how they can help with safer deployments and A/B testing. But I'm concerned about the potential for the codebase to become messy with numerous conditional checks over time.

Do you use tools like LaunchDarkly, or have you built custom solutions? How do you ensure that old flags are cleaned up and that the system remains maintainable?

Would love to hear how your team handles this, especially in larger projects.