Feature Flags

Feature Flags

A Python feature flag system for SaaS applications. Supports boolean flags, percentage rollouts, user targeting rules, A/B testing with deterministic variant assignment, kill switches, and scheduled activations — all built on Python's standard library.

Features

• Boolean flags — Simple on/off toggles for any feature
• Percentage rollouts — Gradual rollout via deterministic hashing (consistent per user)
• User targeting — Rules matching on user attributes (plan, email, role, etc.)
• A/B testing — Multiple variants with consistent assignment per user
• Kill switch — Emergency one-call disable for any flag
• Scheduled flags — Auto-enable/disable at specified times
• Match operators — eq, neq, contains, in, gt, lt
• Evaluate all — Bootstrap all flags for a user in one call (great for frontends)
• JSON persistence — Save/load flag configs and evaluation logs

Requirements

• Python 3.10+
• No external dependencies (stdlib only)

Quick Start

# Start with demo flags
python src/main.py --init-demo

# Custom port
python src/main.py --port 8003

Then try the API:

# List all flags
curl http://localhost:8003/api/flags

# Evaluate a flag for a user
curl -X POST http://localhost:8003/api/flags/new_dashboard/evaluate \
  -H "Content-Type: application/json" \
  -d '{"context": {"user_id": "user_123", "plan": "pro", "email": "user@docs.example.com"}}'

# Kill switch
curl -X POST http://localhost:8003/api/flags/new_dashboard/kill

# Revive
curl -X POST http://localhost:8003/api/flags/new_dashboard/revive

API Endpoints

... continues with setup instructions, usage examples, and more.