LaunchDarkly Pricing 2026: Service Connections, MAU and What a Team Pays
LaunchDarkly pricing in 2026: Developer is free, Foundation bills per service connection and per 1,000 client MAU, no seat fee; a 100,000-user app pays about $883.

LaunchDarkly costs nothing on the Developer plan and, on Foundation, $10 per service connection a month plus $8.33 per 1,000 client-side monthly active users, billed yearly, with no seat fee and no platform fee. A three-service app with 20,000 client users pays about $176.60 a month; the same app at 100,000 users pays about $843. Enterprise is a custom quote.
Those rates were read on LaunchDarkly's pricing page on 17 October 2026. The plan is unusual for developer tooling because the number of engineers never appears on the invoice. What you pay for is the number of server-side SDKs talking to the platform and the number of end users who evaluate flags in a browser or mobile app, so a two-person team serving a consumer app can pay more than a fifty-person team serving an internal tool.
Toolradar data: the feature flag ranking we run evaluated 23 tools, updated 15 October 2026. LaunchDarkly is listed as an alternative to 19 of them, the second most cross-shopped product in the category. The switching question is handled on this site in LaunchDarkly alternatives; this page is the invoice itself.
The plans at a glance
| Plan | Price | What is included | Who it fits |
|---|---|---|---|
| Developer | Free, no card | Unlimited seats, 5 service connections, 1,000 client-side MAU, 5,000 AI runs, 100,000 experimentation MAU, 1 project, 3 environments, 14-day data retention | Side projects and a single-service pilot |
| Foundation | $10 per service connection and $8.33 per 1,000 client-side MAU a month, billed yearly | First 5 service connections free each month, 5,000 AI runs then $5 per 1,000, 5,000 session replays, 5,000 errors, 10M logs, 10M traces, unlimited projects and environments, 30-day retention | Growing teams that want usage pricing without a sales call |
| Enterprise | Custom | Everything in Foundation plus advanced targeting, custom roles, workflows, approvals, custom AI run volumes, 100-plus days of retention | Companies that need approvals and audit at scale |
| Guardian add-on | Priced separately, Enterprise only | Release monitoring, guardrail metrics, automatic rollback, exposure insights | Teams that want guarded rollouts past the included number |
A feature flag platform, for a developer, is a rules engine that decides at runtime which users see which code path, with SDKs that evaluate those rules in the server and the client. LaunchDarkly bills the two halves of that definition separately, which is the whole point of the next section.
What a service connection is, and why it multiplies
LaunchDarkly defines a service connection as one server-side SDK connected to one environment for a month. That sentence hides a multiplication. A backend with four services, each running a server SDK in production and staging, is eight service connections, not four. Add a development environment and it is twelve. Foundation includes the first five every month, so that twelve-connection setup bills seven of them, $70 a month before any client users.
Two consequences for how you architect. First, a monolith is cheap here: one service in three environments is three connections, inside the free five. Second, a microservice estate pays for every service that evaluates flags server-side, so teams on LaunchDarkly tend to route flag evaluation through a shared relay or a small number of edge services rather than putting the SDK in every container. The Developer plan's cap of five connections is a hard limit, not a billable allowance, so a microservice pilot outgrows the free plan before it has any real users.
What a client-side MAU is, and when it dominates
Each unique client-side user or device that evaluates a flag in a month is one MAU, billed per 1,000 on Foundation at the rate in the table above. A consumer web or mobile app pays almost entirely on this meter. At 20,000 users the MAU line is $166.60 a month; at 100,000 users it is $833, and at a million users it is $8,330 before any volume negotiation, which is the point at which Enterprise and a custom contract become the real price list.
Server-side evaluation sidesteps the meter. If your frontend asks your own API which variant to render and the API holds the server SDK, LaunchDarkly sees one service connection and zero client MAU. That architecture is common on LaunchDarkly for exactly this reason, and it is why a B2B SaaS with 5,000 logged-in users can pay a tenth of what a consumer app with the same engineering team pays. The trade-off is that client-side targeting on device attributes, and the session replay and error tracking Foundation now bundles, need the client SDK.
Four costed scenarios
Each total multiplies the yearly-billed Foundation rates. None is a quote, and taxes are excluded.
| Scenario, one month | Service connections | Client-side MAU | Monthly cost |
|---|---|---|---|
| Side project, one service, one environment, 800 users | 1 (inside the free 5) | 800 (inside the free 1,000) | $0 on Developer |
| B2B SaaS, two services in production and staging, 5,000 users evaluating server-side | 4 (inside the free 5) | 0 | $0 on Foundation, the trial converts to a zero bill until usage grows |
| Three-service app, production and staging, 20,000 client users | 6, so 1 billable | 20,000 | $176.60 |
| Five-service app, production and staging, 100,000 client users | 10, so 5 billable | 100,000 | $883 |
The second row is the surprise. Foundation has no base fee, so a team that stays inside five service connections and evaluates flags server-side can sit on the paid plan with its unlimited projects, 30-day retention and bundled observability and pay nothing until the architecture changes. LaunchDarkly's own trial copy says it: 14-day trial, then pay as you go, no platform or seat fees.
The fourth row is the one buyers quote. Five billable connections cost $50 and the MAU line is $833, so a mid-sized consumer app pays $883 a month, or $10,596 a year, on list. That is the figure the feature flag platforms ranking compares against, and it is where the alternatives below start to look different.
How LaunchDarkly compares with three alternatives
| Tool | Entry paid price | What the bill counts | At the 100,000-user, five-service scenario |
|---|---|---|---|
| LaunchDarkly Foundation | $10 per service connection, $8.33 per 1,000 client MAU, yearly | Server SDK connections and client users | $883 |
| Statsig Pro | $150/mo | Metered events, 5M included then $0.05 per 1,000; flag checks unlimited | $150 if events stay inside 5M, flags themselves free |
| ConfigCat Pro | $110/mo | Config downloads, 25M included; seats and connections unlimited | $110, capped at 100 flags and 3 environments |
| Unleash Pay-As-You-Go | $75 per seat a month, 5-seat minimum self-hosted | Seats, plus 53M API requests included then $5 per million | $375 for five seats, user count irrelevant |
Competitor rates were read on each vendor's pricing page on 17 October 2026: Statsig pricing, ConfigCat pricing, Unleash pricing and PostHog pricing, which gives 1 million flag requests a month free with no minimum spend and is not in the table because its bill counts requests rather than users.
The pattern: LaunchDarkly is the only one of the four whose price grows with your end users. Statsig and ConfigCat grow with events and downloads, which a team can cap with client-side caching; Unleash grows with engineers. For a consumer app, LaunchDarkly is the most expensive of the four at scale; for a server-side B2B product inside five connections, it is free. What the extra money buys is the targeting, approval and workflow depth on Enterprise and the Guardian rollback layer, which none of the three alternatives sells as a packaged product.
The costs the pricing page does not total for you
AI runs. Foundation includes 5,000 a month and bills $5 per 1,000 after that. A team using the AI configuration features in production will exceed the allowance quickly, and the overage is a separate line.
Observability allowances. Foundation bundles 5,000 session replays, 5,000 errors, 10 million logs and 10 million traces a month. Usage beyond that bills as you go, and the page does not publish the unit rates, so a team moving its error tracking into LaunchDarkly should confirm them before the move.
Data retention. Developer keeps 14 days, Foundation 30, Enterprise 100 or more. An audit or an experiment that needs a quarter of history is an Enterprise feature, not a Foundation add-on.
Guardian. Guarded rollouts, the automatic rollback when a release degrades a metric, come with an included number on paid plans; after that the feature needs Enterprise with the Guardian add-on, which is priced separately and not listed.
Yearly billing. The per-connection and per-1,000-MAU rates above are the yearly-billed figures LaunchDarkly shows. The page does not publish a monthly-billed rate, so a team that wants month-to-month should check the current price on LaunchDarkly's site before assuming the same numbers.
Who should pay for LaunchDarkly
Pay for it when the flag system is part of how you release, not a toggle: staged rollouts across many services, approvals before a flag changes in production, targeting rules an auditor will read. Those teams buy Enterprise, and the Foundation rates are a stepping stone, not the destination. Flags of that kind sit next to the pipeline, so the CI/CD pricing comparison and the best CI/CD tools ranking are the neighbouring decisions.
Stay on Developer or a zero-usage Foundation when you have one or two services and evaluate server-side. Leave for a request-metered tool when you run a consumer app with hundreds of thousands of users and need flags more than you need approval workflows; at that scale the MAU line is the whole bill, and the feature flag platforms ranking prices the field, with the AI-assisted release side covered in best AI for DevOps. The Toolradar feature flag statistics page tracks the category's pricing mix if you want the wider numbers.
How we priced this
We read LaunchDarkly's pricing page, including its definitions of service connection and client-side MAU and its plan feature lists, plus the Statsig, ConfigCat, Unleash and PostHog pricing pages, on 17 October 2026. The four scenarios multiply the yearly-billed Foundation rates: billable connections above the free five at the per-connection rate, and client MAU at the per-1,000 rate. Cross-shopping figures come from Toolradar, the software directory Dupple runs. We run no test lab, so nothing here is a measurement of the product in use. Louis Corneloup, founder of Toolradar and Dupple, edited the verdict; editorial rules are on the Devshot about page.
FAQ
How much does LaunchDarkly cost in 2026?
The Developer plan is free with unlimited seats, 5 service connections and 1,000 client-side monthly active users. Foundation charges $10 per service connection a month above the first five and $8.33 per 1,000 client-side MAU, billed yearly, with no seat or platform fee. Enterprise is a custom quote, and the Guardian rollback add-on is priced separately on top of Enterprise. Rates were read on LaunchDarkly's pricing page on 17 October 2026.
Does LaunchDarkly charge per seat?
No. Every plan, including the free Developer tier, has unlimited seats. The two meters are service connections (one server-side SDK in one environment for a month) and client-side monthly active users (one unique user or device evaluating flags). A fifty-engineer team serving an internal tool from two services can pay nothing, while a two-engineer team serving 100,000 consumers pays about $833 a month on the MAU line alone.
Is there a free LaunchDarkly plan?
Yes. Developer is free forever, needs no credit card, and includes 5 service connections, 1,000 client-side MAU, 5,000 AI runs, 100,000 experimentation MAU, one project with three environments and 14 days of data retention. The limits are hard caps rather than billable allowances, so a microservice pilot or any consumer app outgrows it quickly, at which point Foundation's 14-day trial converts to pay-as-you-go.
What is a LaunchDarkly service connection?
One server-side SDK connected to one environment for one month. Four services with the SDK in production and staging are eight service connections. Foundation includes five a month and bills each additional one at the per-connection rate, so the count rises with every service and every environment that evaluates flags server-side. Routing evaluation through a relay or a small set of services keeps the number down.
How much does LaunchDarkly cost for 100,000 users?
About $883 a month on Foundation for a five-service app running in production and staging: five billable connections plus 100,000 client-side MAU at the per-1,000 rate. Evaluating flags server-side instead, so the client never calls LaunchDarkly, would drop the MAU line to zero and the bill to the connection charge alone. At that user count Statsig Pro costs $150 with unlimited flag checks and ConfigCat Pro $110, with their own caps on events and flags.
Is LaunchDarkly worth it compared with Statsig, ConfigCat or Unleash?
For release governance, yes: approvals, custom roles, workflows and the Guardian rollback layer are what Enterprise sells, and the three alternatives do not package them the same way. For flags alone on a consumer app, no: LaunchDarkly's bill grows with end users, while Statsig and ConfigCat grow with events and downloads and Unleash grows with engineers at $75 per seat. The break-even is the architecture, since a server-side product inside five connections pays LaunchDarkly nothing.
Cite this: Devshot, "LaunchDarkly Pricing 2026: Service Connections, MAU and What a Team Pays", October 2026.
Devshot is the daily brief on tools and changes developers act on. Join at devshot.co.
Some offers on this page may be paid placements or contain affiliate links.
Devshot: the daily dev & AI brief
Free daily newsletter, read in 5 minutes.
Subscribe free