FeatureSignals Documentation

FeatureSignals is an open-source, AI-powered feature flag management platform built for modern engineering teams. Ship features safely with targeted rollouts, run data-driven A/B experiments, clean stale flags automatically with AI, and recover from incidents in seconds — all without vendor lock-in or surprise bills.

Why FeatureSignals?

• AI-Powered Lifecycle Management — AI scans your codebase, identifies stale flags, analyzes usage patterns, and generates cleanup pull requests. Human reviews, AI does the grunt work.
• Self-hosted & free — Run on your own infrastructure. No per-seat pricing. Apache-2.0 licensed.
• Real-time updates — Server-Sent Events push flag changes to SDKs instantly.
• A/B experimentation — Built-in variant assignment with consistent hashing and impression tracking.
• Toggle categories — Classify flags as release, experiment, ops, or permission for lifecycle-aware management.
• Multi-environment — Manage dev, staging, and production from a single Flag Engine.
• Enterprise-ready — RBAC, audit logging, approval workflows, SSO (SAML/OIDC), SCIM, IP allowlists.
• Compliance-first — SOC 2 controls, GDPR, HIPAA, ISO 27001, CCPA/CPRA, DORA, and CSA STAR readiness.
• SDKs for every stack — Go, Node.js, Python, Java, .NET, Ruby, React, and Vue with OpenFeature support.
• Edge-ready — Relay proxy for low-latency flag evaluation at the edge.

Quick Links

What do you want to do?	Go to
Get up and running in 5 minutes	Quickstart
Understand the core concepts	Feature Flags
Learn about AI capabilities	AI Flag Cleanup
Integrate with your application	SDKs
Explore the REST API	API Reference
Deploy to production	Self-Hosting
Set up A/B experiments	A/B Experimentation
Review security & compliance	Security Overview
See how we compare	vs LaunchDarkly

Architecture at a Glance

┌─────────────┐     ┌─────────────────┐     ┌──────────────┐
│ Flag Engine  │────▶│   Go API Server  │◀───│  PostgreSQL   │
│  (Next.js)   │     │   (chi router)   │     │  (data store) │
└─────────────┘     └────────┬────────┘     └──────────────┘
                             │
              ┌──────────────┼──────────────┐
              │              │              │
         ┌────▼───┐   ┌─────▼────┐   ┌────▼────┐
         │  SDKs   │   │  Relay    │   │  SSE     │
         │Go/Node/ │   │  Proxy    │   │ Stream   │
         │Py/Java/ │   │  (edge)   │   │(real-time)│
         │React    │   └──────────┘   └─────────┘
         └─────────┘

AI Strategy

FeatureSignals uses AI where it solves real problems — not as a gimmick:

1. AI Flag Cleanup (Critical) — Scan codebase, identify stale flags, analyze usage, generate cleanup PR. Human reviews and approves.
2. AI Anomaly Detection (High) — Monitor evaluation patterns, detect anomalies, alert with root cause analysis. Human investigates.
3. AI Incident Response (High) — Correlate changes with errors, suggest rollback. Human approves.

Human-in-the-loop for AI: AI suggests, human approves. Never autonomous production changes.

License

FeatureSignals is licensed under Apache-2.0.