A self-hosted cloud native feature flag solution for managing feature toggles and A/B testing experiments.

GO Feature Flag is an open-source, self-hosted feature flag solution designed for managing feature toggles and A/B testing experiments. It allows developers to enable or disable functionality without deploying new code, supporting a wide range of environments from backends to frontend applications and mobile apps. This tool facilitates a faster iteration cycle by decoupling feature releases from code deployments.
The software is primarily deployed as a server via a relay proxy that provides a central API for flag evaluation, which enables support for multiple programming languages. For projects written exclusively in Golang, it can be used as a native Go module that performs flag evaluation directly within the application code. Configuration files are stored in formats such as JSON, TOML, or YAML and can be hosted in various locations including S3, Kubernetes, or via HTTP.
The architecture relies on the relay proxy to enable multi-language support, allowing the system to integrate with various SDKs across the OpenFeature ecosystem. It is designed for developers and DevOps teams that require a cloud-native approach to continuous delivery and experimentation while maintaining full control over their infrastructure. To evaluate flags accurately, the system requires a unique targeting key for each user, which is passed within an evaluation context alongside other user attributes like email or admin status.
This tool provides a lightweight infrastructure for managing application variants and feature releases across a full technology stack, ensuring that feature toggling remains vendor-neutral and scalable.
Manage feature flags and gradual rollouts to control how and when new features reach end users.
Manage feature flags and run A/B tests using a warehouse native approach to product analytics and experimentation.
Join our newsletter to get shiny new open source software delivered to your inbox. Unsubscribe anytime.