ConfigCat
Hosted feature flags with unlimited seats, metered on config downloads rather than users
ConfigCat is a hosted feature flag and remote configuration service from Budapest that lets development teams turn features on and off, target user segments, and run percentage rollouts without redeploying, using open source SDKs that download a config JSON file from a global or EU-only CDN and evaluate every flag locally in the application; every plan has unlimited seats and unlimited monthly active users, the meter is config JSON downloads, and pricing runs from a Forever Free plan to Pro at $110 a month billed annually.
Overview
ConfigCat is a feature flag service and nothing more, which needs saying plainly because it sits on a shelf labeled product analytics. It does not collect events, draw funnels, compute retention, record sessions, or calculate experiment results. What it does is store flags and settings, publish them as a config JSON file to a content delivery network, and let an SDK inside your application decide locally whether a given user sees a feature. When you want to know whether that feature moved a number, ConfigCat sends the flag evaluation to Amplitude, Mixpanel, Google Analytics, or Twilio Segment and the measurement happens there.
The commercial idea that separates it from the rest of the category is the meter. Seats, monthly active users, service connections, and flag reads are unlimited on every plan including the free one; the only thing counted is how many times your applications download the config JSON, plus the bytes that download moves. Forever Free covers 5 million downloads a month with no card. Pro is $110 a month billed annually, or $120 month to month, for 25 million downloads and 100 flags. Smart is $325 a month annually or $360 monthly for 250 million downloads and what the pricing page calls unlimited flags. A ten-person engineering team pays the same as a two-person one, which is the opposite of Unleash's $75 per seat and the same instinct as LaunchDarkly's unlimited seats, without LaunchDarkly's per-user meter.
The catch is where ConfigCat puts its limits instead. The free plan allows 10 feature flags per product across 2 products and 2 environments, with 2 segments and 4 targeting rules per flag, so it is a real evaluation tier rather than something a growing product lives on. Pro raises flags to 100 and environments to 3, which a team with a normal development, staging, and production split and a busy roadmap will outgrow. In exchange, the security features other vendors charge a tier for, including SSO, SAML, SCIM user provisioning, and two factor authentication, appear on the pricing table for every plan including free, and the zombie flag tooling for finding stale flags is also on every plan.
The company is small, bootstrapped, and Hungarian. The vendor's own history says three former colleagues launched the service in 2018 and spent their own money through a first year with no paying customers; the Terms of Service name the operator as ConfigCat Kft, registered in Budapest. Third-party profiles list about 13 employees and no outside funding. The product has not stood still for it: the changelog shows an approval flow with scheduled changes in August 2026, a rewritten .NET SDK in July, a Startup Program in June, SCIM in 2025, and a self-hosted ConfigCat Proxy in 2024, and the company holds ISO/IEC 27001:2022 certification. The tradeoff is the one you would expect from a vendor this size: fewer integrations and no statistics, but published pricing, flat billing, and an on-premise edition sold at a published price rather than an open-ended enterprise quote.
Best for
Small and mid-sized development teams who want a hosted feature flag service with flat, predictable billing, where adding engineers, product managers, or users never changes the price, and who need SSO and a clean ISO 27001 answer without paying for an enterprise tier to get them.
Not the right fit for
- Anyone shopping for product analytics; ConfigCat has no event tracking, funnels, retention, or session replay of its own and hands flag data to Amplitude, Mixpanel, Google Analytics, or Segment for any measurement at all.
- Teams who want experiment statistics in the flag tool; percentage options split traffic consistently, but there is no significance calculation, no sequential testing, and no metrics engine, so GrowthBook, Statsig, or LaunchDarkly's experimentation are the places to look.
- Products with many long-lived flags on a small budget; the free plan stops at 10 flags per product and Pro at 100, so a flag-heavy codebase lands on Smart at $325 a month sooner than the download volume would suggest.
- Teams who require an open source platform they can fork and run for free; the SDKs are open source but the dashboard and service are not, and the on-premise edition is a commercial license tied to the Dedicated column at $4,500 a month billed annually.
- Buyers whose security review hinges on SOC 2; ConfigCat publishes ISO/IEC 27001:2022 and GDPR compliance and does not publish a SOC 2 report on its pricing page or Trust Center.
How it works
- 1
You sign up with an email address, create a product, and get environments and configs inside it. A config is a collection of feature flags and settings, where a flag is a boolean and a setting can also return a string, integer, or double. Each environment has its own SDK key and its own flag values, so the same flag can be on in staging and off in production.
- 2
Your application installs one of ConfigCat's open source SDKs, which cover JavaScript, React, Angular, Vue, Node.js, Bun, Deno, Cloudflare Workers, .NET, Java, Go, PHP and Laravel, Python, Ruby, Elixir, Android, iOS and macOS, Kotlin Multiplatform, Flutter and Dart, React Native, C++, Rust, Unity, Unreal Engine, and a Chromium extension. The SDK downloads the config JSON from ConfigCat's CDN, which is served through Cloudflare's edge cache, and caches it locally on a polling frequency you set anywhere from one second to several hours.
- 3
Evaluation happens entirely inside your application. When code asks for a flag value, the SDK applies the targeting rules to a User Object you pass in, containing an identifier, email, country, or any custom attribute, and returns the answer from the cached file without uploading any user information. Targeting rules combine user conditions, prerequisite flag conditions, and segment conditions, and percentage options split users into consistent groups for phased rollouts and A/B splits. Reads are never counted, only downloads, so a flag checked a million times between two downloads costs one download.
- 4
Changes made on the dashboard, through the Public Management API, through the Terraform provider, or from an AI editor via the ConfigCat MCP server are published to the CDN and picked up on the next poll. Webhooks notify your systems when flags change, the audit log records who changed what, and from August 2026 environments can require change requests to be approved by a second person, with changes scheduled to apply at a set time. Teams that need a component inside their own network can run the ConfigCat Proxy, a small Go application that caches config files, evaluates flags server side, and streams changes over Server-Sent Events or gRPC.
Feature breakdown
23 features in 4 modulesFlags, settings, and targeting
The core of the product, with limits that tighten on the cheaper plans.- Feature flags and typed settings
- Boolean flags for on and off decisions plus settings that return strings, integers, or doubles, so copy, thresholds, and timeouts can change without a release. Text settings can be set to reject invalid JSON before saving, added in July 2026.
- Targeting rules
- Rules match on any attribute in the User Object, such as email, country, plan, or a custom field, and serve a specific value when they match. Free allows 4 rules per flag and Pro 8, with no stated cap from Smart upward.
- Percentage options
- Split users into consistent groups by percentage for phased rollouts, canary releases, and A/B splits. The same user gets the same answer on every platform because evaluation logic is identical across SDKs, according to the vendor.
- Segments
- Reusable user groups referenced by many flags. Free includes 2 segments, Pro 3 per product, and Smart and above effectively unlimited, with a technical limit of 500 per product on Smart.
- Prerequisite flags
- A targeting rule can depend on the value of another flag, so a feature can require a parent feature to be on without duplicating its rules.
- Predefined variations
- Since February 2026 a flag can offer a fixed list of allowed values rather than free-form input, which stops a typo in the dashboard from becoming a production value.
Delivery architecture and the ConfigCat Proxy
Why flag checks cost nothing on the meter and add no latency.- Local evaluation in every SDK
- SDKs download the config JSON, cache it, and evaluate in process without uploading user data, so there is no network call per flag check and a ConfigCat outage leaves the last cached values in place.
- Global and EU-only CDN
- Config files are served through Cloudflare's edge cache with geo-located load balancing by default, or published only to EU nodes when an organization admin switches Data Governance to EU Only.
- ConfigCat Proxy
- A self-hosted Go application, released in April 2024, that caches config files close to your services, can fall back to an external cache if the CDN is unreachable, evaluates flags server side so frontend clients never see the full config, and streams changes over Server-Sent Events and gRPC.
- OpenFeature support
- OpenFeature providers for languages including React, Angular, NestJS, Kotlin, Swift, PHP, Python, Ruby, and Rust, and the Proxy supports the OpenFeature Remote Evaluation Protocol, so application code can be written against the vendor-neutral standard.
- Configurable polling
- You choose how often SDKs download the config, from once a second to once every few hours, which is the single lever that decides both how fast a change reaches users and how many downloads you are billed for.
Governance, security, and tech debt
Unusually generous at the low end, with approvals arriving only in August 2026.- Unlimited seats with permission groups
- Every plan allows unlimited team members. Free has one permission group, Pro two, and Smart and above unlimited, which is where finer-grained control over who can touch production starts.
- SSO, SAML, SCIM, and 2FA on every plan
- The pricing table lists two factor authentication, SSO through Google, Entra ID, GitHub, Bitbucket, and Microsoft, SAML 2.0 with your own identity provider, and SCIM user provisioning across all plans including free, subject to verifying your email domain.
- Approval flow and scheduled changes
- Released August 27, 2026: environments can require a change request approved by a second member before publishing, and approved changes can be scheduled for a specific time. It works on V2 configs only, and the pricing page does not state which plans include it.
- Audit log
- A record of who changed what and when, with an optional mandatory reason for each change. Retention is 7 days on Free, 35 days on Pro and Smart, and 2 years on Enterprise and Dedicated.
- Zombie flags report
- A dashboard page, a scheduled email, and an API that list flags not changed within a period you define, filtered by environment and tag, so stale flags surface for removal. Listed as tech debt tools on every plan.
- Code references and the CLI
- The ConfigCat CLI and scanners for GitHub Actions, CircleCI, Bitrise, and Bitbucket Pipelines find where each flag is used in source code and upload the references, and flag flags that no longer exist on the dashboard.
Integrations and automation
Where any measurement happens, since ConfigCat does none itself.- Analytics integrations
- Amplitude, Mixpanel, Google Analytics, and Twilio Segment receive flag evaluations or change events, which is how a percentage split becomes an experiment someone can read. ConfigCat's docs also explain passing variation IDs to analytics tools yourself.
- Monitoring and chat
- Flag changes appear as Datadog events tagged by product, config, and environment, and Slack receives change notifications with configurable filters.
- Project management and IDEs
- Jira Cloud, monday.com, and Trello can toggle linked flags, and plugins for VS Code and JetBrains IDEs manage flags from the editor.
- Public Management API, Terraform, and MCP server
- Everything on the dashboard is reachable through the Public Management API, a Terraform provider manages flags, webhooks, and integrations as code, and an MCP server lets Cursor or Claude Desktop create and change flags.
- Webhooks and workflow automation
- Webhooks per environment, 1 on Free, 3 on Pro, and effectively unlimited above that, plus Zapier and Zoho Flow connectors for no-code automation.
- LaunchDarkly import
- Since April 2025 ConfigCat can import flags from LaunchDarkly, and its docs include a migration guide, which is who the vendor is plainly trying to win customers from.
Use cases
4 documentedEight-person SaaS team paying per seat for flags
Every new engineer and product manager adds to the flag vendor's bill, so access gets rationed and people ask in Slack whether a feature is on.
ConfigCat has unlimited seats on every plan, so the whole team gets dashboard access with SSO, and the bill is Pro at $110 a month billed annually regardless of headcount.
Mobile app waiting on app store review
A timeout value and a piece of onboarding copy are wrong in the shipped build, and fixing them means another release and another review.
Both become ConfigCat settings that the SDK downloads on its polling interval, so they change from the dashboard with no new binary, and flag checks are served from the local cache without counting against the plan.
Small team running its first A/B test
Product wants to try two pricing page layouts and already uses Mixpanel, but has no experimentation platform and no budget for one.
A ConfigCat flag with percentage options splits users consistently and the Mixpanel integration logs which variation each user saw. The statistics happen in Mixpanel, and the team has to accept that ConfigCat will not tell them who won.
European company with a data protection review
Legal wants assurance that user attributes used for targeting do not leave the company, and that flag data stays in the EU.
SDKs evaluate locally without uploading user data, Data Governance is set to EU Only so config files are published only to EU nodes, and the ConfigCat Proxy runs server-side evaluation so the full rule set never reaches browsers.
Pricing
from $0 (Forever Free), then $110 per month billed annually or $120 billed monthly (Pro)Freemium subscription with flat plan prices, metered on config JSON downloads and network traffic, with unlimited seats, monthly active users, and service connections on every plan. Extra quota is sold in fixed blocks, and a Dedicated plan covers private cloud or on-premise deployment.
| Plan | Price | Includes |
|---|---|---|
| Forever Free | $0 per month |
The download allowance is generous; the 10-flag cap is what ends the free ride. |
| Pro | $110 per month billed annually ($120 billed monthly) |
Self-serve with a card. Three environments is tight if you run more than development, staging, and production. |
| Smart | $325 per month billed annually ($360 billed monthly) |
The last self-service plan, and the one most growing products end up on because of flag and environment counts rather than traffic. |
| Enterprise | $900 per month billed annually ($1,000 billed monthly) |
|
| Dedicated | $4,500 per month, billed annually only |
The on-premise page describes a flat price under a standard agreement and asks you to request a quote, so confirm the on-premise figure directly. |
Add-ons
- Extra quota on Free ($45 per month): +5 million downloads and +20 GB traffic, bought in advance; $70 if charged as overage.
- Extra quota on Pro ($95 per month): +25 million downloads and +100 GB traffic, bought in advance; $140 if charged as overage.
- Extra quota on Smart ($270 per month): +250 million downloads and +1 TB traffic, bought in advance; $410 if charged as overage.
- Extra quota on Enterprise ($760 per month): +1 billion downloads and +4 TB traffic, bought in advance; $1,200 if charged as overage.
- Extra quota on Dedicated ($3,300 per month): +6 billion downloads and +24 TB traffic, bought in advance; $5,200 if charged as overage.
- Software escrow (Quoted): Enterprise and Dedicated only; access to the source code if ConfigCat ceases operating.
Billing notes
- Prices as published September 2026 on configcat.com/pricing, where the default view is billed annually. Paying month to month costs $120 instead of $110 on Pro, $360 instead of $325 on Smart, and $1,000 instead of $900 on Enterprise; the vendor rounds that to saving up to 10 percent.
- The meter is config JSON downloads plus network traffic, counted per calendar month in UTC and including HTTP 304 Not Modified responses. Flag reads, seats, monthly active users, and connected instances are not counted.
- Going over does not switch your flags off. The billing policy sets excess usage charges above the prepaid block price, for example $140 per extra 25 million downloads on Pro against $95 if bought in advance, and the Terms of Service give you 30 days after notice to reduce usage, upgrade, or pay before the account can be disabled.
- Unused add-on quota does not roll over. Prices apply to card payments; wire transfer is by request with a $200 administrative fee, and prices through AWS Marketplace may differ.
- The billing policy lists surcharges for non-standard requests: 10 percent to disable automatic renewal, $1,200 to sign a customer-provided NDA or data processing agreement, and $600 for a security questionnaire over 20 questions.
- Startups accepted to the Startup Program get 50 percent off Pro or Smart annual plans in year one and 25 percent in year two, and student and teacher pricing exists. USD prices rose 1 percent and EUR prices 10 percent for new customers in February 2025.
Value assessment: ConfigCat's pricing is easy to model once you stop thinking about users. A backend service polling once a minute uses about 43,000 downloads a month per running instance, so ten instances sit comfortably inside the free 5 million. Browser and mobile clients are the variable: if each of 100,000 monthly users triggers ten downloads, that is a million downloads, still inside Free, and 1 million users on the same pattern fits Pro's 25 million with room to spare. For most small businesses the download meter will never bind; the flag and environment limits will. Ten flags per product on Free and 100 on Pro push a flag-heavy team to Smart at $325 a month well before traffic does. Against the neighbors the arithmetic is simple: Flagsmith's Start-Up is $40 a month billed yearly for three seats, cheaper for a small team that does not need SSO; Unleash is $75 per seat, so ConfigCat Pro is cheaper for any team of two or more; LaunchDarkly's free tier gives more flags but meters client-side monthly active users above 1,000, which ConfigCat never does. The real value is SSO, SAML, and SCIM at no extra cost, which elsewhere costs $250 a month or an enterprise contract.
Entry price history
Entry price first recorded on 26 September 2026. It will be re-checked when this category next comes round the rotation.
| Checked | Entry price | What happened |
|---|---|---|
| 26 September 2026 | $110 | First recorded |
Dates are when SaaSTracker checked the vendor’s published pricing, not when the vendor changed it. See the price changes log for movements across the whole database.
Strengths & limitations
Strengths
- Unlimited seats, monthly active users, service connections, and flag reads on every plan, so neither headcount nor user growth changes the bill.
- SSO, SAML 2.0, SCIM provisioning, and two factor authentication appear on the pricing table for every plan including free, where Flagsmith gates SSO at $250 a month and LaunchDarkly gates SCIM behind Enterprise.
- Local evaluation in every SDK means no per-check network call, no user data uploaded for targeting, and cached values that keep working if the CDN is unreachable.
- Broad SDK coverage, from Cloudflare Workers, Bun, and Deno to Unity, Unreal Engine, C++, and Elixir, with all SDKs open source on GitHub.
- A choice between a global CDN and an EU-only CDN, ISO/IEC 27001:2022 certification, and a Trust Center with the audit report, penetration test sample, subprocessor list, and a DPA.
- Fully published pricing including overage rates, prepaid blocks, a published on-premise and private cloud tier, and a changelog showing steady shipping through 2026.
Limitations
- No analytics and no experiment statistics. Percentage options split traffic and integrations send evaluations to Amplitude, Mixpanel, Google Analytics, or Segment, but ConfigCat never tells you whether a variation won.
- Tight flag and environment limits on the cheaper plans: 10 flags and 2 environments per product on Free, 100 flags and 3 environments on Pro, so structure rather than traffic forces the upgrade to $325 a month.
- Approval flow and scheduled changes only shipped on August 27, 2026, work only on V2 configs, and the pricing page does not say which plans include them.
- The service itself is closed source; the only self-hosted path is the commercial on-premise edition priced with the Dedicated plan at $4,500 a month billed annually, not a free community build like Flagsmith or Unleash.
- No SOC 2 report is published, only ISO/IEC 27001:2022 and GDPR, which some US procurement teams will treat as a gap.
- Frontend SDKs download the whole config file, targeting rules included, so anything sensitive in a rule is visible in the browser unless you run the ConfigCat Proxy for server-side evaluation.
Head-to-head comparisons
3 alternativesConfigCat vs Flagsmith
from $0 (Free), then $40 per month billed yearly (Start-Up)Flagsmith is cheaper for a small team that needs nothing but flags: unlimited flags even on its free tier and Start-Up at $40 a month billed yearly for three seats, against ConfigCat Pro at $110 with a 100-flag cap. ConfigCat wins once the team grows or security enters the picture, with unlimited seats on every plan and SSO, SAML, and SCIM included free where Flagsmith starts them at $250 a month. Flagsmith also has a free open source self-hosted build; ConfigCat's on-premise edition is a commercial $4,500 a month tier.
Full ConfigCat vs Flagsmith comparisonConfigCat vs LaunchDarkly
from $0 (Developer), then pay-as-you-go from $8.33 per 1,000 client-side monthly active users per month billed yearly (Foundation)Both give unlimited seats, but they meter different things: LaunchDarkly counts client-side monthly active users above its free 1,000 and bills $8.33 per thousand on Foundation, while ConfigCat counts only config downloads and a busy consumer app can stay on Pro at $110 a month. LaunchDarkly brings experimentation statistics, session replay, and observability that ConfigCat does not attempt, plus unlimited flags on its free tier. Choose ConfigCat for predictable flag billing and SSO without an enterprise contract; choose LaunchDarkly if you want measurement and release control from one vendor.
Full ConfigCat vs LaunchDarkly comparisonConfigCat vs Unleash
from $75 per seat per month (Pay-As-You-Go cloud)Both evaluate flags locally so user data stays in your application, and both offer EU hosting. Unleash charges $75 per seat with no free hosted tier, so a three-person team pays $225 a month; ConfigCat Pro is $110 billed annually for any number of seats, and its free plan is real. Unleash has the longer governance track record, an open source build you can self-host for free, and approvals that are well established, while ConfigCat's approval flow is weeks old. Small teams usually land cheaper on ConfigCat; regulated teams that need a free self-hosted option look at Unleash.
Full ConfigCat vs Unleash comparisonPairs well with
7 categoriesConfigCat is a product analytics tool, and a team running it still has 7 neighbouring categories to staff. The pick in each is Dataddo for customer data platforms, Product Fruits for product onboarding & PLG, ChartMogul for retention & churn prevention, tawk.to for customer support platforms, Microsoft Clarity for CRO & experimentation, Creem for subscription billing and Tally for forms & surveys, each the leading tool in that category on published criteria.
- DataddoCustomer Data Platforms · Ease of Use
- Because
- ConfigCat lists the service itself is closed source among its limits. Dataddo adds predictable per-flow pricing with unlimited users and sources.
- Benefit
- Events get collected once and sent everywhere, instead of a tag per destination.
- Product FruitsProduct Onboarding & PLG · Best Value
- Because
- ConfigCat is built for product analytics, not onboarding. Product Fruits adds certifications well beyond what a 27-person company usually carries.
- Benefit
- New users reach the thing the product is for without a call, and the drop-off shows up in the data.
- ChartMogulRetention & Churn Prevention · Best Value
- Because
- ConfigCat lists tight flag and environment limits on the cheaper plans among its limits. ChartMogul adds a genuinely free plan up to $10,000 MRR with full metrics.
- Benefit
- Accounts at risk surface while there is still time to call, rather than at renewal.
- tawk.toCustomer Support Platforms · Best Value
- Because
- ConfigCat is built for product analytics, not support. tawk.to adds visitor monitoring and proactive triggers.
- Benefit
- Questions arrive in one queue with a history attached, instead of scattered across personal inboxes.
- Microsoft ClarityCRO & Experimentation · Ease of Use
- Because
- ConfigCat sends buyers elsewhere for experiment statistics in the flag tool. Microsoft Clarity adds free with no traffic limits.
- Benefit
- Page changes get tested rather than argued about, and the winner is measured.
- CreemSubscription Billing · Best Value
- Because
- ConfigCat is built for product analytics, not billing. Creem adds a modern developer surface with a TypeScript SDK and a Next.
- Benefit
- Recurring revenue bills, retries and reconciles itself, including the tax nobody wants to calculate.
- TallyForms & Surveys · Best Value
- Because
- ConfigCat lists no SOC 2 report is published among its limits. Tally adds flat workspace pricing with unlimited collaborators.
- Benefit
- Answers arrive structured and routed, rather than as free text somebody has to read and retype.
Pairings follow the methodology: the leading tool in each complementary category, chosen on published criteria. Placement cannot be purchased.
Full ConfigCat stackImplementation & onboarding
- Setup time
- Under an hour to a first flag in production: sign up with an email, create a flag, install the SDK for your language with the environment's SDK key, and wrap a code path. Choosing polling intervals and, for EU customers, setting the dataGovernance option in every SDK to match the dashboard are the two decisions worth getting right on day one.
- Learning curve
- Low. The model is products, environments, configs, and flags, with targeting rules, segments, and percentage options on top, and the dashboard is usable by product managers as well as developers. The concept that trips people up is that updates arrive on the next poll rather than instantly, which is a tradeoff between speed and download count that you control.
- Onboarding
- Self-serve on Free, Pro, and Smart with a card and no sales call, and no trial is needed because the free plan has the same feature set as the paid ones, according to the vendor. Enterprise and Dedicated add assisted onboarding, 4 hours a month of dedicated support, and help with contracts and forms. Documentation includes per-SDK references, an interactive feature flag tutorial, and a Community Slack where the team answers questions.
- Migration notes
- Coming from LaunchDarkly there is a built-in import tool, released April 2025, plus a migration guide. OpenFeature providers mean code written against that standard can switch to ConfigCat by changing the provider rather than the call sites, and the Public Management API and Terraform provider let flags be recreated as code. Going the other way, flags and rules are exportable through the API; because ConfigCat stores no user event history, there is no analytics data to extract.
Platform, API & security
- Platforms
- JavaScript, React, Angular, and VueNode.js, Bun, Deno, and Cloudflare Workers.NET and ASP.NET CoreJava and Spring BootGoPHP and LaravelPythonRubyElixirAndroid and Kotlin MultiplatformiOS and macOS (Swift)Flutter and DartReact NativeC++ and RustUnity and Unreal EngineChromium extensionOpenFeature providersConfigCat Proxy
- API
- Public Management API covering flags, settings, environments, webhooks, integrations, audit logs, and zombie flags, plus a Terraform provider, an MCP server for AI editors, a CLI, and webhooks per environment. The self-hosted ConfigCat Proxy exposes evaluation endpoints including the OpenFeature Remote Evaluation Protocol, with Server-Sent Events and gRPC streaming.
- Compliance
- ISO/IEC 27001:2022GDPR
- Data residency
- Config files are served from Cloudflare's edge cache on a Global setting by default, or from EU nodes only when an organization admin sets Data Governance to EU Only, which the vendor says keeps the data in the EU. The service is run from Hungary under Hungarian law. User attributes used for targeting are evaluated inside your application and are not uploaded. Dedicated customers can place their API and CDN nodes in chosen locations, and the on-premise edition runs entirely on your infrastructure.
- SSO
- Google, Entra ID, GitHub, Bitbucket, and Microsoft sign-in, SAML 2.0 with your own identity provider, and SCIM 2.0 user provisioning, all listed on every plan on the pricing page; SAML and SCIM require verifying your company email domain.
- Security notes
- The Trust Center lists annual third-party audits, a disaster recovery plan, insurance, third-party penetration testing, a bug bounty and vulnerability disclosure program, a subprocessor list, a willingness to sign a DPA, and full customer data deletion, with the ISO certificate, audit report, Statement of Applicability, and CAIQ available. No SOC 2 report is published. Frontend SDKs download the full config file, so sensitive values in targeting rules should be evaluated through the ConfigCat Proxy instead.
Support & resources
- Channels
- Best-effort support and the Community Slack on FreeStandard support by email on Pro and SmartPremium SLA support with a meaningful answer by the end of the next business day and 4 hours of dedicated support a month on Enterprise and DedicatedFree demo sessions on request
- Documentation
- Documentation at configcat.com/docs covering every SDK, targeting, segments, polling modes, data governance, zombie flags, the Proxy, the MCP server, SSO and SCIM, plan limits, and a dated product changelog, plus Public Management API docs at api.configcat.com.
- Community
- A public Community Slack where the ConfigCat team answers questions, open source SDKs on GitHub, and a public roadmap. The vendor cites a ThoughtWorks Technology Radar mention on its site.
Company
- Founded
- 2018
- Headquarters
- Budapest, Hungary
- Ownership
- Bootstrapped, privately held (legal entity ConfigCat Kft)
- Employees
- Approximately 13 (Tracxn, May 2026); the vendor reported over 30 in July 2022
- Funding
- Bootstrapped; the vendor describes founders funding the first year from their own savings, and Tracxn lists no funding rounds.
Timeline
- 2018Three former colleagues launch ConfigCat in Hungary as a hosted remote configuration and feature flag service, funding it themselves through a first year without paying customers.
- 2019The first paying customer signs up, buying what was then the $200 top plan, according to the vendor's own history.
- 2022The vendor reports roughly 200 paying customers, a team of over 30, and ISO certification, after growing 5 times in 2020 and 3 times in 2021 by its own account.
- 2024OpenFeature providers ship in January, Config V2 in April, the ConfigCat Proxy in April, and analytics integrations for Mixpanel, Amplitude, Google Analytics, and Twilio Segment in May and June.
- 2025USD prices rise 1 percent and EUR prices 10 percent for new customers in February, a LaunchDarkly import tool arrives in April, SCIM user provisioning in May, and Proxy profiles with OpenFeature Remote Evaluation Protocol support in August.
- 2026Predefined variations ship in February, AUD and GBP billing in January, a Startup Program with 50 percent off in June, a rewritten .NET SDK in July, and approval flow with scheduled changes on August 27.
Integrations
- Amplitude
- Mixpanel
- Google Analytics
- Twilio Segment
- Datadog
- Slack
- Jira Cloud
- monday.com
- Trello
- GitHub Actions
- GitLab
- Bitbucket Pipelines
- Bitrise
- CircleCI
- Terraform
- VS Code
- JetBrains IDEs
- Zapier
- Zoho Flow
- OpenFeature
- MCP server
Frequently asked questions
13 questionsWhat is ConfigCat?
ConfigCat is a hosted feature flag and remote configuration service. You create flags and settings on its dashboard, its open source SDKs download them as a config JSON file from a CDN, and your application evaluates them locally to decide which users see which features. It is run by ConfigCat Kft in Budapest and priced on config downloads, with unlimited seats on every plan.
How much does ConfigCat cost?
Forever Free is $0 with no card. Pro is $110 a month billed annually or $120 monthly. Smart is $325 a month annually or $360 monthly. Enterprise is $900 a month annually or $1,000 monthly. Dedicated, which covers private cloud or on-premise deployment, is $4,500 a month billed annually. Prices are also published in EUR, GBP, and AUD.
Is ConfigCat a product analytics tool?
No. It sits in this category because feature flags are bought alongside analytics, but ConfigCat does not track events, build funnels, measure retention, or record sessions. It splits users with targeting rules and percentage options and sends flag evaluations to Amplitude, Mixpanel, Google Analytics, or Twilio Segment, where the analysis happens.
Can ConfigCat run A/B tests?
It can serve them but not score them. Percentage options divide users into consistent groups, and the analytics integrations record which variation each user saw. There is no significance calculation or experiment report in ConfigCat, so you read the result in your analytics tool, or use GrowthBook, Statsig, or LaunchDarkly if you want the statistics built in.
What counts as a config JSON download?
Every time an SDK fetches the config file from ConfigCat's CDN, including requests answered with 304 Not Modified. Reading flag values from the local cache is not counted, so the bill depends on how many instances and clients you run and how often they poll, not on how often your code checks a flag or how many users you have.
What happens if I go over my plan's limit?
Your flags keep being served. ConfigCat notifies you, and under its Terms of Service you have 30 days to reduce usage, upgrade, or pay excess usage charges, which are higher than buying extra quota in advance: on Pro, an extra 25 million downloads and 100 GB costs $95 prepaid or $140 as overage. If you ignore the notice, the account can be disabled.
Can I self-host ConfigCat?
Only commercially. The on-premise edition ships as Docker containers for the dashboard, API, and CDN nodes with upgrades and 4 hours a month of support, and the pricing table places it in the Dedicated column at $4,500 a month billed annually, while the on-premise page asks you to request a quote. The service is not open source, only the SDKs. The ConfigCat Proxy runs in your infrastructure but still depends on ConfigCat's CDN for its data.
Does ConfigCat keep data in the EU?
It can. Setting Data Governance to EU Only publishes your config files only to EU CDN nodes, and the vendor says your data then never leaves the EU. User attributes used for targeting are evaluated inside your application rather than sent to ConfigCat. The company is Hungarian, holds ISO/IEC 27001:2022, and will sign a DPA.
Is SSO really included on the free plan?
According to the pricing table, yes: two factor authentication, SSO through Google, Entra ID, GitHub, Bitbucket, and Microsoft, SAML, and SCIM provisioning are listed across all plans. SAML and SCIM require verifying your company's email domain first. That is unusual in a category where SSO often costs $250 a month or an enterprise contract.
Who owns ConfigCat?
ConfigCat Kft, a company registered in Budapest, Hungary. It was started in 2018 by three former colleagues and is bootstrapped, with the vendor describing the founders living off their savings through the first year. Tracxn lists no outside funding and about 13 employees. Customer names on the site include Speechify and, in the company's history, H&M.
What should I pair ConfigCat with?
ConfigCat is a product analytics tool, and a team running it still has 7 neighbouring categories to staff. The pick in each is Dataddo for customer data platforms, Product Fruits for product onboarding & PLG, ChartMogul for retention & churn prevention, tawk.to for customer support platforms, Microsoft Clarity for CRO & experimentation, Creem for subscription billing and Tally for forms & surveys, each the leading tool in that category on published criteria.
What do I pair with ConfigCat for Customer Data Platforms?
Dataddo. ConfigCat lists the service itself is closed source among its limits. Dataddo adds predictable per-flow pricing with unlimited users and sources.
Is ConfigCat enough on its own?
For product analytics, yes. ConfigCat does not cover product onboarding & PLG, retention & churn prevention and customer support platforms, and its own profile says so; Product Fruits, ChartMogul and tawk.to are the tools that fill those gaps.
Editorial verdict
ConfigCat is the flag service to pick if you want billing that ignores your team size and your user count. Unlimited seats and monthly active users on every plan, a download meter that local evaluation keeps small, SSO, SAML, and SCIM without an enterprise tier, an EU-only CDN, and ISO 27001 add up to a strong offer from a small, bootstrapped Budapest company that keeps shipping. Pro at $110 a month billed annually is fair for a team of any size, and for teams of three or more it undercuts Unleash's per-seat model outright. Be clear about what it is not. It is not product analytics despite the shelf it sits on, and it will not tell you whether an experiment won; pair it with Mixpanel, Amplitude, or PostHog for that. Watch the structural limits more than the traffic ones, because 10 flags on Free and 100 on Pro push a busy codebase to Smart at $325 before downloads ever matter, and if you need a free self-hosted build or long-established approval workflows, Flagsmith and Unleash are the better fits.
Written by the SaaSTracker editorial team. Awards, when shown, are judged against the published criteria in our methodology.