Based on the knowledge sources provided, GrowthBook primarily uses feature flags and SDKs for implementing experiment variations, but it also offers a visual editor for certain types of experiments. Here's a summary of the options: 1. Feature Flags: You define the feature flags in the GrowthBook interface, but the actual code for the variations needs to be implemented in your codebase where the GrowthBook SDK is integrated. The SDK then evaluates these flags to determine which variation to show. 2. Visual Editor: GrowthBook offers a visual editor for creating simple experiments without requiring code changes. This is part of their Pro plan. As mentioned in the [Running experiments with GrowthBook and GA4](https://docs.growthbook.io/guide/GA4-google-analytics#running-experiments-with-growthbook-and-ga4) documentation: "We do have a visual editor for creating experiments as part of our Pro plan. Our visual editor is meant for simple experiments. Experiments that are more complex are best created by writing code with feature flags." 3. Inline Experiments: These allow you to define experiment conditions and settings directly in your code, which are then processed by the GrowthBook SDK. This approach doesn't require declaring features up-front in the GrowthBook interface. It's important to note that GrowthBook does not currently offer a way to write variant code directly in their interface and then deploy it to the client's website automatically. The implementation of the variations typically needs to be done in the client's codebase, with GrowthBook controlling which variation is shown through feature flags or the visual editor for simpler changes. If you need more detailed information about these options, I recommend checking the GrowthBook documentation or reaching out to their support team for specific implementation guidance.