At a Glance:
Featurevisor is an open-source feature management platform that lets teams manage feature flags, A/B test variations, and remote configuration variables declaratively within a Git repository, with build workflows generating static datafiles for CDN delivery and SDK-based evaluation.
Overview:
Featurevisor is an open-source, Git-based feature management tool for managing feature flags, A/B testing variations, and remote configuration. It works by having users define attributes, segments, and features declaratively in a Git repository. A CI/CD pipeline builds these definitions into static JSON datafiles, which are then uploaded to a CDN or custom server. Applications fetch these datafiles and use Featurevisor’s SDKs to evaluate feature flags, variations, and variables at runtime against user-provided context. The project supports JavaScript (Node.js and browser), React, React Native, Go, Python, PHP, Ruby, Java, Swift, Next.js, and Vue.js.
Key Decision Points:
Git-native workflow: All feature configurations are stored as files in a Git repository, allowing for version control, pull requests, and code review processes.
Build-and-upload model: Feature definitions are built into static JSON datafiles that must be uploaded to a CDN or custom server; this requires setting up a CI/CD pipeline like GitHub Actions.
SDK-based evaluation: Applications evaluate features against runtime context using provided SDKs, which support flag toggles, string variations for A/B tests, and key/value variables for remote configuration.
Declarative building blocks: The project uses three core concepts — attributes (conditions), segments (reusable targeting rules), and features (definitions combining flags, variations, and variables) — to structure feature management.
Core Features:
Feature flags: Define on/off toggle rules evaluated in applications using SDKs against provided context.
A/B test variations: Configure string-based variation values for experiments, with targeting controlled through segments and attributes.
Remote configuration variables: Manage key/value pair variables within feature definitions for runtime configuration adjustments.
Segments: Create reusable targeting conditions from attribute definitions to consistently apply rules across multiple features.
Datafile generation: Build the declarative project files into static JSON datafiles for delivery via CDN or custom server, decoupling feature evaluation from a live service.
Use Cases:
Developers can decouple application deployments from feature releases using feature flags managed in a Git repository.
Teams running A/B and multivariate experiments can define test variations declaratively and evaluate them client-side through SDKs.
Projects requiring remote configuration can manage key/value variables in Git and apply changes by rebuilding and uploading updated datafiles, without redeploying applications.
Open-Source Alternative Value:
Featurevisor provides an open-source approach to feature management where all configurations live in a Git repository, giving teams a history of changes through version control. The architecture separates configuration management from evaluation by generating static JSON datafiles that are served from infrastructure users already control — their own CDN or server. SDKs in multiple languages handle the runtime evaluation, making the system embeddable into existing applications without reliance on an external feature flag service for request evaluation.
