More AI code. Less release risk.

What is FeatBit?

FeatBit is an open-source feature flag platform for reducing release risk as AI builds more features faster and introduces more bugs. Teams control each feature's release audience and connect observability signals with flag changes to assess release health throughout the rollout. When problems appear, they can restore previously healthy release settings and audiences, preserving access for healthy users.

Feature flag definition

A feature flag is a runtime control for application behavior. Once integrated, it separates a behavior change from a code deployment. OpenFeature: feature flags (opens in a new tab)

How does FeatBit help control AI-generated changes?

AI produces more features faster, with more bugs to contain. Use FeatBit to validate with QA in production, then target by customer and role, segment, or percentage.

A release health workflow connects metrics from your observability tools with FeatBit flag changes. Continuously monitor feature failures and infrastructure health in real time at every release stage. Use that context to guide expansion or restore previously healthy release settings and audiences when problems appear.

Matching a targeted audience

FeatBit targeting rules match a user only when every condition is met. Combine customer and role attributes to define a specific release audience. FeatBit: targeting rules (opens in a new tab)

Release health signals

Datadog RUM can correlate feature releases with user experience and performance changes once feature flag tracking is configured. Datadog: feature flag tracking (opens in a new tab)

How does feature rollback work?

FeatBit can restore recorded, previously healthy release settings and audiences, manually or through configured external automation. Keep the feature available to the earlier healthy audience while reversing a failing expansion.

Record healthy settings, maintain a tested fallback, and connect monitoring alerts to the recovery workflow through the REST API, CLI, or MCP.

Rollout assignments

The FeatBit CLI accepts percentage allocations for each flag variation. Allocations must total 100%. FeatBit CLI: rollout assignments (opens in a new tab)

External alert integration

A configured New Relic alert can call FeatBit's REST API through a webhook to change rollout settings for a specific user group. FeatBit: New Relic alert integration (opens in a new tab)

Does FeatBit support A/B testing?

Yes. A/B testing supports business validation as one part of FeatBit's release control workflow. Use feature flag variations to form comparison groups, and define exposure and metric events before starting. Bayesian analysis helps assess value and uncertainty; results can remain inconclusive.

Experiment inputs

Flag evaluation events record which variations FeatBit serves to users. Experiment metric events record the interactions or conversions a team chooses to track for business comparisons. FeatBit: experiment data flows (opens in a new tab)

Can I self-host FeatBit?

Yes. Run FeatBit in your own cloud or on-premises infrastructure with Docker Compose or Kubernetes. Your team operates the deployment, its supporting services, and the SDK integration in your application.

SDK evaluation

FeatBit server-side SDKs evaluate flags locally. Client-side SDKs receive flag evaluation results from the server. FeatBit: SDK evaluation models (opens in a new tab)

What is included in the open-source edition?

The free self-hosted Foundation edition includes feature flags, user targeting, segments, SDKs, REST APIs, audit logs, and experimentation.

FeatBit is open core, with most code under the MIT license. SSO, approval and scheduling workflows, and fine-grained access control require a commercial license. Cloud is managed hosting with its own plans.