Blog

Notes from the build.

Technical writing for founders who ship. How Betterflag gets built, real numbers behind our pricing and launch decisions, and hands-on guides for flags, rollouts, and agent-driven releases. No content-marketing filler.

All Architecture Engineering Growth Insights

GrowthAug 24, 20264 min read

LaunchDarkly Audit Logs Are Enterprise-Only. That Is Backwards.

LaunchDarkly gates audit logs, approvals, SCIM, and custom roles behind Enterprise. Why the kill switch's paper trail should not be a sales call, and what Scale at $99.99 includes instead.

GrowthAug 24, 20264 min read

LaunchDarkly Pricing in 2026: What You'll Actually Pay

A teardown of LaunchDarkly pricing: $10 per service connection plus $8.33 to $10 per 1,000 client-side MAUs. Worked examples, a calculator, and the one-meter alternative.

GrowthAug 23, 20264 min read

The Best LaunchDarkly Alternatives in 2026

Honest LaunchDarkly alternatives for 2026: PostHog, Statsig, GrowthBook, Flagsmith, Unleash, ConfigCat, and Betterflag, organized by why you're leaving, not by feature checklists.

GrowthAug 23, 20264 min read

Why LaunchDarkly Bills You for Your Architecture

LaunchDarkly service connections and client-side MAUs price your topology and your success, not your flag usage. Worked math for 25k and 100k MAU, and the one-meter alternative.

GrowthAug 22, 20265 min read

Feature Flag Pricing in 2026: What You'll Actually Pay

A comparison of feature flag pricing in 2026: LaunchDarkly MAUs and service connections, PostHog per request, seat-priced clouds, and one-meter evaluation billing.

EngineeringAug 22, 20264 min read

Feature Flags Over MCP: Give Your Coding Agent a Kill Switch

What happens when your feature flag platform is a first-class MCP server: how agents create flags, stage rollouts, and pull kill switches, and why guardrails and audit trails matter more than dashboards.

EngineeringAug 21, 20264 min read

Shipping a Feature From Claude Code: Flag, Rollout, Kill Switch

A hands-on walkthrough of a full agent-driven release: Claude Code (or Cursor) writes the code, creates the flag over MCP, stages a 10% production rollout, watches evaluation stats, and rolls forward. No dashboard required.

EngineeringAug 21, 20265 min read

Feature Flag Kill Switch: Instant Rollback Without a Redeploy

A kill switch is a feature flag that turns a broken path off in seconds. How it differs from a revert, what to wrap, and how to pull it from a dashboard or an agent.

EngineeringAug 20, 20265 min read

How to Do a Percentage Rollout with Feature Flags

A practical guide to percentage rollouts: sticky assignment, 10% then 50% then 100%, kill criteria, and how gradual rollouts differ from canary deploys.

EngineeringAug 19, 20265 min read

Feature Flags in Next.js App Router

How to use feature flags in Next.js App Router: Server Components, middleware, edge evaluation, caching, and client components without killing performance.

InsightsAug 19, 20269 min read

Why I'm Building Betterflag

Agents made shipping faster than releasing is safe. Feature flags are the control plane for that world, and the current tools were built for a different one.

EngineeringAug 18, 20265 min read

How to Add Feature Flags in React

Add feature flags in React without a flash of the wrong UI: server evaluation, a useFlag hook, targeting, percentage rollouts, and tests. Copy-paste patterns for 2026.

ArchitectureAug 17, 20265 min read

Feature Flags vs Environment Variables: When to Use Each

Environment variables flip at deploy. Feature flags flip per user, at request time. A clear split for secrets, config, remote config, and runtime rollouts.

EngineeringAug 16, 20266 min read

Feature Flag Best Practices That Actually Scale

Practical feature flag best practices: naming, defaults, targeting, percentage rollouts, kill switches, and how to retire flags before they become debt.

EngineeringAug 15, 20266 min read

What Are Feature Flags? A Practical Guide for 2026

Feature flags (also called feature toggles) let you deploy code without releasing it. How they work, the four primitives that matter, and when you actually need them.