Feature Flags
Flux feature flags let you ship code without releasing it. Roll features out gradually, run A/B tests, and target specific users, organizations, or entitlement tiers — all without a deploy.
Creating a Flag
Flags are created from the Flags dashboard or via the API. Each flag has a key (e.g. new-dashboard), a type (boolean or multivariate), and a default value.
Boolean Evaluation
The simplest flag type — on or off.
use flux_sdk::Flux;
let flux = Flux::init("your-api-key");
if flux.flags().is_enabled("new-dashboard", &user)? {
// Render new dashboard
}Python
import flux
flux.("your-api-key")
if flux.flags.("new-dashboard",):
# Render new dashboard
passMultivariate Evaluation
Return one of several string variants — useful for A/B tests or configuration values.
let variant = flux.flags().get_variant("checkout-flow", &user)?;
match variant.as_str() {
"control" => render_current_checkout(),
"variant-a" => render_streamlined_checkout(),
_ => render_current_checkout(),
}Targeting Rules
Targeting rules let you enable a flag for a subset of users or organizations without a full rollout. Rules are evaluated in order; the first match wins.
Common targeting dimensions:
- User attribute — email domain, country, plan tier
- Organization — specific org IDs or org attributes
- Percentage rollout — gradual ramp by user ID hash
- Entitlement — only users on plans that include a given entitlement
Rules are configured from the Flags dashboard with no SDK changes required.
Tenant Overrides
You can override a flag’s value for a specific organization — useful for beta customers or enterprise-specific behavior.
// Check flag with org context (tenant override takes priority)
let enabled = flux.flags().is_enabled_for_org("new-dashboard", &org.)?;Entitlement-Aware Flags
Link a flag to an entitlement so it is automatically enabled only for organizations whose billing plan includes that entitlement.
Configure the entitlement link in the Flags dashboard. The SDK evaluation call is the same — Flux resolves the entitlement check server-side.
Next Steps
- Billing & Entitlements — Define the entitlements that drive flag targeting
- Organizations — Understand how tenant context flows into evaluations
- Observability — Trace flag evaluations and measure experiment impact