Feature flags are not a hard product to understand. A rules engine, a dashboard, and a fast way to get flag values into your code. What is hard to understand is why that costs five figures a year, and why the usual alternative is a config table someone built in a sprint and nobody owns.
We built Flipswitch because that gap shouldn’t exist anymore. The rest of this post is the reasoning.
What changed in the last few years
Building a feature flag platform used to take a big team and a long runway. You needed to design a custom SDK in every language, build a reliable propagation layer, write your own evaluation engine and ship it consistently across six runtimes, and convince customers to lock into your protocol. The cost of getting that far was a moat as much as the product was.
That’s mostly not true anymore. Three things changed:
1. OpenFeature exists. A CNCF-hosted standard SDK interface across every major language, with a working community, real adoption, and the political weight to keep vendors honest. There’s no reason to build another bespoke SDK in 2026.
2. OFREP exists. The OpenFeature Remote Evaluation Protocol is a small, open HTTP API for flag evaluation. It means SDKs can be thin HTTP clients. No evaluation engine to maintain in N languages. No version skew between SDKs. The hardest part of shipping SDKs is now upstream and shared.
3. The infrastructure is commodity. SSE for real-time updates is mature. Edge caching is one config away. Managed Postgres is a utility you rent by the month. None of it is novel.
Together these mean a small team can ship a reliable feature flag platform without the big team and long runway the incumbents needed. It also means the pricing they set back then doesn’t reflect what the service costs to provide today.
What we decided to do
Three decisions followed: stay narrow, build on open standards, and price for what the service actually costs.
Focus on feature flags only. Flags, targeting, and percentage splits for A/B/n tests are in scope. Product analytics and session replay are not. Those tools exist and are good at their jobs, and bundling them tends to produce a worse product across the board. We integrate with the tools you already use rather than competing with them. The roadmap is narrower, the surface area is smaller, the docs are shorter, and the thing we ship works.
OpenFeature native, not OpenFeature compatible. The official Flipswitch SDKs are OpenFeature providers built on top of the official OpenFeature SDKs. There’s no proprietary fork underneath, no Flipswitch-specific API to learn. Switching to us means changing one provider initialization line. Switching away from us is the same. We’d rather compete on infrastructure and pricing than on lock-in, and being honest about that has to start with the SDK choice.
Transparent pricing, real free tier. Flat tiers visible on the website. No per-seat charges. Real-time updates aren’t gated behind a paid plan, because they aren’t expensive for us to provide. The free tier doesn’t expire and doesn’t ask for a credit card. If you can’t size up the cost from the pricing page in 30 seconds, we’ve failed.
What we deliberately didn’t do
We didn’t build a fat SDK, and we didn’t write one from scratch either. The official Flipswitch SDKs are the open source OpenFeature SDKs that already exist, with real-time updates added on top. Targeting rules live on our server, not on every browser and mobile device. Evaluation is a network hop, but every SDK caches, so it isn’t a round trip on every call. In exchange the SDKs stay simple and behave the same in every language. The server-side vs client-side article walks through that decision in detail.
We didn’t add analytics. You can run A/B/n tests with Flipswitch, serving variants by percentage split, but measuring the outcome is a different product. Dashboards, significance tests, and metric pipelines belong in the analytics stack you already have.
Where that leaves you
Building a feature flag platform got cheaper, but the contracts priced for the old cost are still around. You don’t have to choose between those and rolling your own. And because everything runs on OpenFeature, switching to Flipswitch, or away from it, is just a provider change.