Editor Sebastiaan Smits ·
Published October 29, 2026 · Updated October 29, 2026
Teams look past LaunchDarkly when two usage meters make the bill hard to forecast, when approvals and custom roles sit on a quoted Enterprise plan, or when they only want flags. Most move to ConfigCat for flat pricing, to Flagsmith or Unleash to self-host, and to GrowthBook or Statsig when experiments matter.
Short answer: The closest replacements for LaunchDarkly in our feature flag software ranking are ConfigCat, Flagsmith and Unleash. Choose ConfigCat if you want a fixed monthly bill and an official import tool for your LaunchDarkly flags and segments. We took the other 11 from the Best Feature Flag Software in 2026 ranking too, and each is compared with LaunchDarkly alone, noting where staying put is the better call.
Visibility in our rankings can be paid for. Payment moves a vendor's position within the shortlist; it never adds a vendor, and it never changes a word of the review or of this page. The first three alternatives go to challengers, vendors outside the best-known brands in a category, wherever the category has them; the established names follow in their ranked order. How it works: placement disclosure · editorial process.
The Foundation plan has no platform or seat fee but charges $10 per service connection a month and $8.33 per 1,000 client-side monthly active users, billed yearly. Kubernetes, serverless, Python and Ruby setups are quoted separately, because LaunchDarkly says per-connection pricing is not a fair fit for them.
Custom roles, workflows with approvals, scheduled flag changes and SCIM provisioning appear on the Enterprise plan, whose price is quoted. Release monitoring, guardrail metrics and automatic rollback belong to Guardian, a paid add-on priced separately. SSO and SAML are on every plan.
LaunchDarkly's pricing page now lists AgentControl for AI runs, plus session replays, error monitoring, logs and traces as metered items. A team that wants flags alone pays attention to a larger catalog, and Data Export of raw event data is an add-on to select plans.
Best replacement at the top. Each row ends with where we place that product among the 13 we rank.
| # | Alternative | Headquarters | Switch if | Our rank |
|---|---|---|---|---|
| 1 |  ConfigCat | Hungary | you want a fixed monthly bill and an official import tool for your LaunchDarkly flags and segments | #1 |
| 2 |  Flagsmith | United Kingdom | you need the option to run the flag service on your own servers, with an importer to start from | #4 |
| 3 |  Unleash | Norway | you want open-source roots and a documented, phased route off LaunchDarkly | #5 |
| 4 |  GrowthBook | United States | experiments matter as much as flags and your event data already sits in a warehouse | #6 |
| 5 |  DevCycle | United States | you want OpenFeature across the stack and an importer you can read and run yourself | #3 |
| 6 |  Flipt | Not stated | you would rather review flag changes in pull requests than in a vendor dashboard | #12 |
| 7 |  FeatBit | United States | you will host it yourself and want the yearly price to stay flat regardless of traffic | #11 |
| 8 |  PostHog | United States | product analytics is the bigger need and flags are a means to test features inside it | #8 |
| 9 |  Statsig | United States | experiment results and flag rollouts should come from the same event data | #9 |
| 10 |  Harness Feature Management & Experimentation | United States | your releases already run through Harness pipelines and flags should plug into them | #2 |
| 11 |  AWS AppConfig | United States | your workloads already run on AWS and you want flags priced like any other cloud API call | #13 |
| 12 |  Optimizely | United States | you want flags from the same vendor that already sells you experimentation | #10 |
Flat monthly plans where seats, monthly active users and service connections are unlimited on every tier, so growth does not raise the bill. The Forever Free plan holds 10 flags. It publishes an official LaunchDarkly import wizard and lists 19 SDK platforms, though the pages read describe flags and configs rather than experiment statistics.
Switch from LaunchDarkly if you want a fixed monthly bill and an official import tool for your LaunchDarkly flags and segments.
Hosted, private cloud or self-hosted, with the source on GitHub under a BSD-3-Clause licence. Billing counts API requests, and the free plan allows one team member. A built-in LaunchDarkly importer copies projects, flags and segments, though the docs advise checking the result by hand.
Switch from LaunchDarkly if you need the option to run the flag service on your own servers, with an importer to start from.
Built around activation strategies, and the self-hosted open-source core is AGPL-3.0 with no seat limit but one project and two environments. Enterprise adds SSO, change requests and a support SLA. Its docs include a phased migration guide written for LaunchDarkly users.
Switch from LaunchDarkly if you want open-source roots and a documented, phased route off LaunchDarkly.
Experiments read from your own data warehouse, or from a managed one, so exposure data stays where analysts already work. Pricing counts seats instead of traffic, and the free Starter plan includes unlimited flags and experiments, but caps users at three.
Switch from LaunchDarkly if experiments matter as much as flags and your event data already sits in a warehouse.
Every SDK speaks OpenFeature, and an open-source importer script pulls environments, features and variables out of a LaunchDarkly project. The free plan covers 1,000 client-side users. The site announces that DevCycle is now part of Dynatrace, following an asset purchase, so check who your contract would be with.
Switch from LaunchDarkly if you want OpenFeature across the stack and an importer you can read and run yourself.
Flags live in Git, so a change reaches production as a reviewed commit, and the server ships as one binary under GPL-3.0. Pro is a flat monthly price that covers up to 10 instances. Its LaunchDarkly guide begins by swapping your code to OpenFeature SDKs.
Switch from LaunchDarkly if you would rather review flag changes in pull requests than in a vendor dashboard.
MIT-licensed and self-hosted first, with no limits on seats, users, projects or service connections on any self-host plan. Enterprise is a yearly fee, and advanced role-based access costs extra. Experiments with Bayesian metrics appear in its plan table; the page read covered the self-host plans only, not the cloud ones.
Switch from LaunchDarkly if you will host it yourself and want the yearly price to stay flat regardless of traffic.
Flags are one product among analytics, replay, surveys and experiments, billed per request with a million free each month. Pick it when flag data should land in the same tool as behavior data. It is built around product teams, so check your approval and role needs before moving.
Switch from LaunchDarkly if product analytics is the bigger need and flags are a means to test features inside it.
A single event stream feeds flags, experiments, analytics and session replay, and flag and config checks are unlimited on the free tier. Billing is by events, not connections. It leans toward experimentation, so teams that want only toggles pay for analytics they may never open.
Switch from LaunchDarkly if experiment results and flag rollouts should come from the same event data.
Sold as a module of the Harness delivery platform, so rollouts, release monitoring and kill switches connect to pipelines you may already run. Paid pricing sits behind sales. Teams on another CI/CD tool get little from that link, and there is no self-hosting information on the product page.
Switch from LaunchDarkly if your releases already run through Harness pipelines and flags should plug into them.
Read the full Harness Feature Management & Experimentation review · Open vendor site →
An AWS service that bills per configuration request, per configuration received and per experiment hour, with no seat charge in its price list. It validates configurations and can roll back on a CloudWatch alarm. Everything assumes an AWS account and AWS-style tooling.
Switch from LaunchDarkly if your workloads already run on AWS and you want flags priced like any other cloud API call.
Feature management is one product inside a wider Optimizely line-up, with 20+ SDKs and a 99.9 percent uptime SLA stated on the product page. No prices are published, so expect a sales process and a larger vendor relationship than flags alone.
Switch from LaunchDarkly if you want flags from the same vendor that already sells you experimentation.
Stay if your organization has many squads that need custom roles, approval workflows and audit trails across projects, because those controls are part of LaunchDarkly's Enterprise plan. The Developer plan is also a fair holding position: it is free, has unlimited seats and flags, 30 SDKs, five service connections and 1,000 client-side users.
ConfigCat, Flagsmith and DevCycle each read your flags through a LaunchDarkly API token, and Unleash and Flipt publish step-by-step guides. Segments and targeting rules map across with some manual checks. The SDK calls in your code are the larger job, and no vendor automates that part for you.
LaunchDarkly publishes OpenFeature providers for .NET, Cloudflare, Java, Node.js and PHP. Wrapping evaluations in OpenFeature first lets you run both vendors in parallel and switch by changing the provider. Flipt's guide takes this route, and DevCycle builds every SDK on it.
Data Export of flag requests and experiment events needs an add-on, so check whether you hold it before cancelling. Under LaunchDarkly's public terms, paid subscriptions renew for the same term at then-current rates unless canceled before the next billing date, and fees are non-refundable.
Yes, flags can be read through the REST API, for example GET /api/v2/flags/{projectKey} with an access token. The importers from ConfigCat, Flagsmith and DevCycle use that API. Segments and teams may need manual follow-up, and ConfigCat says permissions and code changes are not automated.
Yes, under the public terms of service, paid subscriptions renew for periods equal to the initial term and are charged at then-current rates unless you cancel in the account dashboard before the next billing date. Enterprise order forms can differ, so read your own contract.
Yes, in most cases, because the LaunchDarkly SDK calls are specific to that vendor. OpenFeature providers exist for .NET, Cloudflare, Java, Node.js and PHP, and putting evaluations behind OpenFeature before the move limits what changes at cutover.
Flagsmith, Unleash, FeatBit, Flipt and GrowthBook can be self-hosted, and ConfigCat lists on-premise hosting on its pricing page. Unleash states that its open-source edition has no seat limit. LaunchDarkly's pricing page mentions no self-hosted edition.
Not always, because it depends on which meter drives your bill. A mobile app with many client-side users gains most from flat or seat-based pricing such as ConfigCat or GrowthBook. A team with few connections and users may already pay little on the Developer or Foundation plan.
Raw event data does not move with your flags. LaunchDarkly's Data Export streams to Kinesis, Pub/Sub or Event Hubs, or loads a warehouse in batches, and it is an add-on to select plans. Set up the export before cancelling if you need history.
Each number and claim about LaunchDarkly and its competitors was taken from the vendors' own websites on October 29, 2026; expect them to drift as those sites are edited.
Sebastiaan Smits, Founder, wrote, edited and fact-checked this page.
Founder of The Interregnum and of 2 Steps Ahead, an AI search visibility agency. Editorially responsible for every page here; research and drafting are AI-assisted. Last reviewed October 29, 2026.
How we review.