{"id":"split-api","name":"Split.io (Harness Feature Flags)","homepage":"https://www.split.io","repo_url":"https://github.com/splitio/python-client","category":"devops","subcategories":["feature-flags","experimentation","developer-tools"],"tags":["split","feature-flags","experimentation","a-b-testing","harness","sdk","rest-api"],"what_it_does":"Feature flag and experimentation platform (now part of Harness) with REST API and SDKs for managing feature splits, measuring impact with metrics, and safely rolling out changes.","use_cases":["Feature flag evaluation in agent decision trees to alter behavior based on enabled features","Experiment configuration — creating splits and assigning traffic percentages via API","Reading metric impact of feature rollouts for automated reporting","Progressive delivery automation tied to observability metrics","Kill switch control for agents to disable problematic feature paths"],"not_for":["Simple boolean flags without metrics (use Unleash OSS for cheaper option)","Teams requiring fully self-hosted feature management","Very small teams — pricing can be steep for limited usage","Non-engineering-driven feature management workflows"],"best_when":"You need feature flags tightly coupled with metrics and experimentation data — particularly when you want to measure the impact of each flag change with statistical rigor.","avoid_when":"Your feature flag needs are basic, you need self-hosted control, or you're cost-sensitive — LaunchDarkly and Unleash often provide better value.","alternatives":["launchdarkly-api","unleash-api"],"af_score":79.9,"security_score":null,"reliability_score":null,"package_type":"mcp_server","discovery_source":["github"],"priority":"low","status":"evaluated","version_evaluated":"current","last_evaluated":"2026-03-01T09:50:06.247443+00:00","performance":{"latency_p50_ms":25,"latency_p99_ms":120,"uptime_sla_percent":99.9,"rate_limits":"Admin API rate limited; SDK connections are streaming and long-lived","data_source":"llm_estimated","measured_on":null}}