Flagsmith Review (2026)

What Flagsmith actually is

Flagsmith is open-source feature flags and remote config. Flags are the primary job - it also does segmentation and basic A/B testing, but experimentation is not a statistical-engine strength. If you’ve seen the name “Bullet Train,” that’s the same product under its original name.

What makes Flagsmith worth a serious look isn’t a headline feature. It’s the shape of the company. It’s the David in a field of giants - bootstrapped, roughly 12 people, based in London, founded in 2018, running on around $1.5M in revenue. It didn’t raise a $200M round, it isn’t owned by OpenAI, and it publishes its full pricing on a page you can read without talking to sales. In this category, all three of those are unusual.

Two things it does that the giants don’t

It’s genuinely open source, and genuinely self-hostable. The core platform is licensed BSD-3-Clause, and “core functionality stays open, always.” The open-source build includes flags, remote config, segmentation, basic A/B testing, 15+ SDKs, and real self-hosting via Docker, Kubernetes/Helm, on-prem and air-gapped. That’s the thing LaunchDarkly and Statsig structurally can’t offer, and it’s a cleaner self-host story than PostHog’s officially-unsupported one.

Its pricing is the most transparent in the category. Every core tier is published with a number next to it. No “contact sales” to find out what Start-Up costs. In a space where LaunchDarkly’s real contract prices are a mystery until you negotiate, that honesty is worth something on its own.

Pricing

This is Flagsmith’s strongest card, so here it is in full.

Tier	Annual	Monthly	Requests / seats
Free	$0	$0	50k requests/mo, 1 member, unlimited flags
Start-Up	$40/mo	$45/mo	1M requests/mo, 3 members, A/B testing
Scale-Up	$250/mo	$300/mo	5M+ requests/mo, 5 members (up to 20 at $50 each)
Enterprise	Contact sales	-	5M+, 20+ members, self-host, custom SLA

At $40/mo annual, Start-Up is the cheapest paid entry of any tool in this comparison. The billing unit is requests, which is clean and easy to reason about - no MAU trap, no highest-volume-context-kind surprise. Every tier gets unlimited flags, environments, identities and segments, so you’re never rationing the core object. Open-source and nonprofit discounts are available on request. And if you self-host, the OSS build is free outright - you’re paying for cloud convenience, not for the software.

Self-hosting: real, but real work

Let’s be honest about the trade, because “free and open source” hides a cost. The OSS self-host is genuinely free and genuinely capable. But running it in production is real DevOps. Flagsmith is a Django and Postgres stack, which is more resource-intensive than the Node (Unleash) or Go (Flipt) alternatives. Third-party comparisons call it “tricky to get right” - you’re managing container orchestration, a database, upgrades, scaling and ongoing maintenance.

So the “free” self-host quickly costs engineering time instead of license dollars. For a team with DevOps capacity that wants no lock-in and full data control, that’s a great trade. For a small team without that capacity, the paid cloud tiers - which are cheap anyway - are usually the better call.

One caveat worth flagging: the exact list of features gated behind the paid Enterprise license isn’t fully published. It’s governance and management features layered on the open core, likely advanced RBAC, SAML and audit at scale, but we couldn’t verify the precise list from public sources. Check with Flagsmith directly before assuming a specific enterprise feature is in the open build.

Flagsmith versus the field

The clean framing: Flagsmith trades scale for independence and honesty. It has the highest G2 rating of the four at 4.8, though from only 37 reviews - a fraction of LaunchDarkly’s 681, so read it as “small and loved” rather than “battle-tested at every scale.” Against LaunchDarkly, Flagsmith is cheaper, open source and self-hostable, but has fewer SDKs (15+ vs ~38), no dedicated edge story, and a smaller ecosystem. Against Statsig and PostHog, it’s flags-first with only basic A/B testing - if you need a real statistical experimentation engine, it’s not the tool.

The smaller team is the honest risk and the honest upside at once. Smaller support surface and ecosystem, yes - but no acquisition roadmap risk, and because it’s BSD-3 open source, the code outlives the company and you’re never locked in.

Should you use it?

Use Flagsmith if you want open-source feature flags you can genuinely self-host, you value transparent published pricing and no lock-in, and you don’t need heavyweight experimentation. For that brief, it’s the standout choice here.

Don’t use Flagsmith if you need a real statistical A/B engine (Statsig or PostHog), you want the deepest flag tooling and broadest SDK and edge coverage (LaunchDarkly), or you want to self-host but have no DevOps capacity to run a Django stack.

Bottom line: Flagsmith is the refreshingly straightforward option - open, cheap, transparent, and independent. It won’t out-feature the giants, and it doesn’t try to. If open-source flags without lock-in is what you’re after, this is the one to start with.

---

Pricing and features verified against flagsmith.com on 23 July 2026. Feature-flag pricing changes often - we re-verify regularly. Funding figures for Flagsmith conflict across public sources; the exact Enterprise-only feature list was not fully verifiable and is noted as such.