Split

Feature flagging and experimentation platform that decouples release control from code deployment. Teams manage gradual rollouts, target users by segment or percentage, and measure feature impact on business and engineering metrics. SDKs run evaluation locally so targeting data stays in your application.

Split is now part of Harness as Feature Management and Experimentation. Trusted by enterprises such as Experian, Imperfect Foods, and Skillshare for production releases. Differentiates with local flag evaluation for data privacy, streaming updates that push changes in milliseconds, and release monitoring that ties each flag to engineering KPIs and alerts on issues. Global infrastructure serves billions of evaluations daily with SDKs for JavaScript, React, Java, Ruby, iOS, Android, and other platforms, plus Open Feature provider support.

Key features:

• Feature flag management with targeting, segments, and percentage-based rollout plans
• Release monitoring that correlates flag exposure with performance and user behavior, with alerts tied to specific flags
• Experimentation and A/B testing with real-time analytics engine
• Local SDK evaluation with streaming updates and local caching for resilience
• Integrations with analytics, monitoring, and messaging tools

Engineering teams use it for safe continuous delivery, kill switches to turn off features in seconds, and gradual rollouts to limit blast radius. Product and data teams run experiments and measure impact without involving data science. QA integrates flags into smoke, regression, and E2E testing workflows.