Flagsmith

Bootstrapped, 100% open-source feature flags and remote config — built by two engineers in London for teams that need to self-host.

Feature Flags & Remote Config 4.6 / 5 Free OSS · Cloud from $45/mo Updated May 2026
🟡 AMBER — Watching · Best-in-class for self-host, small-vendor watchpoint

Quick Verdict

Flagsmith is the credible, genuinely open-source alternative to LaunchDarkly — and unlike most "open-core" rivals, the core is permissively licensed and the self-hosted version has no artificial usage limits. Co-founded in 2018 in London by Ben Rometsch (CEO) and Kyle Johnson (CTO) as a spin-out from Solid State Group, a London digital agency Ben co-founded in 2002, the product was originally called Bullet Train (the legal entity remains Bullet Train Ltd) and emerged because the agency's own engineers needed feature flags they could self-host for client work. For Indian startups — particularly fintechs under RBI data-residency rules, ABDM-bound healthtechs, and DPDPA-bound enterprises — being able to deploy the entire Flagsmith stack inside an AWS Mumbai VPC for free is a structural advantage that no enterprise SaaS rival can match.

The watchpoint is vendor scale. Flagsmith is bootstrapped — no Series A. Crunchbase and PitchBook list total disclosed funding at roughly $130K–$229K of angel money; the founders have publicly chosen to skip VC. 2024 revenue was about $1.5M on a 12-person team per GetLatka. That's a real, profitable, ethically-run business — but it's a small vendor in a category that just got compressed: in September 2025 OpenAI acquired Statsig for $1.1B, LaunchDarkly continues to scale enterprise revenue, and PostHog has folded free feature flags into a broader product-analytics suite. The open-source core insulates customers from outright vendor failure, but procurement teams at scale should still ask the small-vendor questions.

Quick facts: Founded 2018, London (86–90 Paul Street). Co-founders Ben Rometsch + Kyle Johnson, both ex-Solid State Group. Legal entity Bullet Train Ltd. Bootstrapped — no Series A. ~12 employees, $1.5M revenue 2024. Founding member of OpenFeature (CNCF open standard). 100% open-source core. Cloud Free $0 / Start-Up $45/mo / Enterprise contact-sales. Self-hosted unlimited and free.

Open-Source Credibility
5.0
India Self-Host Fit
4.8
Developer Experience
4.6
Cost Efficiency vs LaunchDarkly
4.9
Vendor Scale
3.1

What is Flagsmith?

Flagsmith is a feature flag and remote configuration service that lets engineering and product teams decouple code deployment from feature release. You ship hidden code to production behind a flag, then turn the flag on for specific users, a percentage cohort, an environment, or a segment defined by traits — all from a dashboard, no redeploy. The same primitives drive canary releases, kill switches, geo-rollouts, paid-tier gating, A/B and multivariate tests, and remote-configuring app strings or JSON.

Origin story. Ben Rometsch co-founded Solid State Group, a London product-engineering agency, in 2002. By the mid-2010s the agency was shipping client products and Kyle Johnson — whose first professional job was at Solid State Group — built an internal feature-flag tool called Bullet Train to help the agency's developers ship client features incrementally. The agency open-sourced Bullet Train, the GitHub repo started getting external traction, and in 2018 Ben and Kyle spun the project out as a standalone company. The product was renamed Flagsmith in 2020 (the legal entity stayed Bullet Train Ltd, which is why Crunchbase still files under that name). The Financial Times' GitHub org still has an archived bullet-train-frontend repo redirecting to the current Flagsmith repo — a fossil of the rename.

Funding and posture. Flagsmith is bootstrapped. PitchBook lists total disclosed funding at roughly $229K; Crunchbase shows ~$130K. There is no Series A, no growth round, no acquisition. Ben has discussed publicly on the Craft of Open Source podcast and elsewhere that the founders chose product simplicity and profitability over hype-cycle growth. GetLatka reported $1.5M revenue in 2024 (up from $1.1M in 2023) with a ~12-person team. That's a deliberately small, profitable open-source business — closer to Sidekiq or Plausible than to LaunchDarkly.

OpenFeature. Flagsmith is a founding member of OpenFeature, the CNCF-incubated vendor-neutral feature-flag standard. They've shipped OpenFeature provider SDKs across major languages. The practical implication for buyers: if you instrument your application against the OpenFeature SDK rather than Flagsmith's proprietary SDK, you can swap to LaunchDarkly, Unleash, or another OpenFeature-compatible backend later with minimal application code change. That's a uniquely buyer-friendly stance in a category historically defined by lock-in.

Capabilities

Feature Toggles

Boolean and multivariate flags. Turn features on/off across web, iOS, Android, and server-side in milliseconds without redeploy. Per-environment isolation (dev/staging/production), audit log, and change history.

Remote Configuration

Beyond true/false: serve strings, integers, and JSON values keyed by flag. Useful for tuning pricing tiers, UI copy, API endpoints, or fraud thresholds without an app-store release.

Segments & Targeting

Define user cohorts by traits (email domain, region, app version, paid tier). Roll out by percentage, traits, or explicit user IDs. Real-time updates without SDK redeploy.

Self-Hosted & Edge

Official Docker images, Helm charts, Terraform modules. Deploy to AWS Mumbai, GCP Mumbai, or your data centre with no per-flag, per-MAU, or per-request licensing limits.

OpenFeature Native

Founding-member status of OpenFeature. Use the vendor-neutral SDK if you want optionality, or Flagsmith's native SDK if you want extra features. Avoid the lock-in trap.

A/B Testing (basic)

Variation assignment and exposure tracking. Honest framing: the statistical engine is functional but not on par with GrowthBook or the late Statsig — pair Flagsmith with a dedicated analytics tool if conversion analytics is the primary use-case.

Pricing (2026, Verified)

Cloud pricing verified at flagsmith.com/pricing in May 2026. ALL-IN figures convert at $1 ≈ ₹84 and add 18% IGST applicable to overseas SaaS for Indian business buyers.

Free (Cloud)

$0/mo
₹0/mo
  • ✅ 50,000 API requests / month
  • ✅ 1 team member
  • ✅ 1 project · unlimited flags
  • ✅ Community Slack support

Enterprise (Cloud)

Contact
Typically ₹50K–₹3L /mo
  • ✅ 5M+ API requests / month
  • ✅ 5–20 seats (more on request)
  • ✅ SSO, RBAC, SLA, priority support
  • ✅ Private Cloud single-tenant option

Self-Hosted (OSS)

$0
Software ₹0 · Infra ₹15K–50K /mo
  • ✅ Unlimited requests, flags, users
  • ✅ Run on AWS Mumbai / GCP Mumbai / on-prem
  • ✅ Full source available on GitHub
  • ⚠️ No SLA, no vendor support
  • ⚠️ Needs DevOps capacity for upkeep

Pricing-schema note (pattern #18 check): Unlike ConfigCat's flat-tier and LaunchDarkly's MAU-based pricing, Flagsmith cloud is request-metered (API calls/month) with a per-seat soft cap. There is no per-flag charge. Self-hosted Enterprise is licensed by seats and projects, not by requests. Annual billing yields a modest discount on Start-Up and Enterprise. Verify your expected request volume before committing — a busy mobile app polling flags every minute can blow through 1M requests/month quickly; switch the SDKs to local-evaluation mode to avoid overage.

When Flagsmith Is the Right Call

  1. Indian fintechs and NBFCs with RBI data-residency obligations. Self-host the full stack inside an AWS Mumbai or GCP Mumbai VPC. No user-targeting data, no IP, no email-domain trait ever leaves Indian borders. The OSS licence is permissive; you don't pay for the privilege.
  2. ABDM-bound healthtechs and DPDPA-bound enterprises. Same self-host argument applies to ABDM consent-manager integrations and post-Aug-2023 DPDPA-Act enterprise procurement. Flagsmith's OSS posture eliminates the cross-border data-transfer hurdle.
  3. Teams escaping LaunchDarkly MAU bills. LaunchDarkly's per-MAU pricing scales brutally as a B2C app grows. Flagsmith cloud at $45 + overage, or self-hosted at zero, is typically 5–20× cheaper at scale. The OpenFeature SDK lets you migrate gradually.
  4. Bootstrapped SaaS and DevTools startups. The free cloud tier (50K requests) covers a meaningful early-stage product; the OSS version covers everything else. No "free trial expires in 14 days" panic.
  5. Teams that genuinely care about open-source ethics. Unlike "open-core" rivals that gate enterprise features behind a proprietary licence, Flagsmith's core is permissively licensed and the GitHub repo is the canonical product, not a marketing artefact.
  6. Long-term technical-strategy buyers. OpenFeature membership means you're not locked in. Even if Flagsmith is acquired, shuts down, or pivots, your application code remains portable.

When Flagsmith Is the Wrong Call

Conversion-analytics-led experimentation: if your primary use-case is statistically rigorous A/B testing tied to revenue or activation metrics, Flagsmith's A/B engine is honest but basic — pair it with PostHog or GrowthBook for the stats layer, or evaluate GrowthBook standalone. Large-enterprise RBAC and audit needs: if you need fine-grained role hierarchies, SAML SSO with SCIM, and detailed compliance audit packages out-of-the-box, LaunchDarkly remains the safe enterprise default — though Flagsmith Enterprise has closed much of the gap. Teams without DevOps capacity that still want to self-host: running Flagsmith OSS in production is not difficult, but it does require someone to own Postgres + Redis + the Flagsmith container, backups, and upgrades. If you don't have that capacity, pay for cloud Start-Up instead. INR-billing-mandatory procurement: billing is in USD with no Indian entity. CFOs requiring rupee invoices and GST input credit will need to route through a reseller or build the procurement workaround. Single-vendor "one throat to choke" enterprise preference: a 12-person bootstrapped vendor is the wrong answer if your procurement team's primary risk filter is vendor size.

Pros and Cons

Pros

  • Genuinely open-source core — not "open-core" theatre
  • OpenFeature founding member → zero lock-in
  • Unlimited self-hosted use, including on AWS Mumbai for RBI/DPDPA
  • 5–20× cheaper than LaunchDarkly at scale
  • Bootstrapped → no VC pressure to extract from customers
  • Profitable, ethically run, transparent founders
  • Docker + Helm + Terraform tooling for self-host

Cons

  • Small vendor (~12 people) — procurement may flag
  • A/B testing analytics are basic vs GrowthBook / late Statsig
  • UI functional, not as polished as LaunchDarkly's
  • No INR billing or Indian entity
  • Self-host requires DevOps capacity to operate
  • Enterprise feature gap narrower than it was, but still real

Bottom line for Indian buyers

If you are an Indian fintech, healthtech, or DPDPA-bound enterprise that needs feature flags and needs the targeting data to stay inside Indian infrastructure, Flagsmith self-hosted on AWS Mumbai is the obvious answer — there is no equally credible alternative at a comparable price point. The OpenFeature standard means you're not locked in if Flagsmith ever changes. The watchpoint is vendor scale: at 12 people, bootstrapped, in a category where OpenAI just paid $1.1B for Statsig in September 2025, Flagsmith is a small ship in a consolidating sea. The OSS escape hatch makes that risk acceptable. Pair Flagsmith with PostHog or GrowthBook if statistical A/B analytics is the primary use-case rather than release engineering.

Book a Free Call