Feature Flags

Feature flags let you control your application’s behavior from GrowthBook, without deploying new code. Target specific users, gradually roll out changes, or run A/B tests on the client or the server. Feature flags decouple deploys from releases. Release code more frequently and with less risk by launching to a subset of users, ramping up gradually, or turning any change into an experiment.

​

Getting Started

To use feature flags, set up an SDK Connection, then explore the docs below.

​

Learn the Basics

• Fundamentals — Keys, feature types, default values, and publishing
• Environments — Control which features are enabled per environment (dev, staging, production)
• Constants — Reusable named values you reference from flag values
• Configs — Typed, validated, composable object shapes you can back JSON flags with (Enterprise)

​

Set Up Targeting and Rules

• Rules — How rules work, choosing a rule type, and configuring forced values, rollouts, experiments, and scheduling
• Targeting Conditions — Define who sees a feature using attributes, conditions, and saved groups
• Safe Rollouts — Release features with automatic guardrail monitoring
• Ramp Schedules — Progressively increase traffic on a rule over a defined timeline
• Prerequisites — Gate features based on the state of other features

​

Manage Flags at Scale

• Publishing & Approval Flows — Drafts, revisions, merge conflicts, and requiring approvals for features and saved groups
• Stale Feature Detection — Identify flags that may no longer be in use
• Code References — Find where flags are used in your codebase
• Feature Diagnostics — Inspect feature evaluations in production
• Custom Hooks — Extend GrowthBook with custom validation logic (self-hosted Enterprise)