featflags docs
Everything you need to put a flag in your code, roll it out safely and measure what it changes. If you’re new, start with Getting started: it takes you from sign-in to your first flag in a few minutes.
Start here
SDKs
Every SDK evaluates flags in your own process, so your users’ attributes never leave it and the same user gets the same answer in every language.
Features
- ExperimentsSplit part of a flag's traffic between variants, measure the metrics you choose, and see which variant did better, without sending us who your users are.
- Release templates and guarded releasesWrite a rollout once, such as staff first, then 10%, 50% and everyone, use it on any flag, and let your metrics pause it when the new variant does harm.
- Approvals, scheduling, triggers and webhooksRequire a second pair of eyes in production, schedule a change for Monday morning, let your alerting turn a flag off, and tell Slack or your own systems about every change.
- Code referencesSee where your code reads each flag, and know when a finished flag is safe to delete, from a scan your CI runs on every push.
- Single sign-onSign your people in with your own identity provider, over SAML or OpenID Connect, with new colleagues joining on their first sign-in.
- Audit logOne log of who did what across your organisation, to search, download as JSON Lines or CSV, and stream to Datadog, Splunk or your own endpoint as it happens.
Tools
- AI assistants (MCP)Connect Claude, Cursor, VS Code or another AI assistant to featflags, and ask it to look up, create, roll out and clean up flags for you, with your access and your approvals.
- CLIThe featflags command for your terminal and CI, for flags, history, key lists, code references, experiment keys, exports and the LaunchDarkly importer.
- TerraformManage projects, environments, SDK keys, flags, targeting rules, segments, release templates, triggers and webhooks as code.
- Importing from LaunchDarklyCopy a LaunchDarkly project into featflags, keep every user in the variant they had, and check every evaluation against LaunchDarkly before you switch.
- API referenceAuthenticate with an API token, call the main endpoints for projects, flags, targeting, lists, history and keys, and handle errors and rate limits.
Reference
- Security and privacyWhat your SDKs send featflags and what they don't, what browsers and apps can see, how experiment data stays pseudonymous, and how access is controlled.
- SDK specificationThe contract every featflags SDK implements, from the flags document to contexts, evaluation, bucketing, membership lists, analytics, experiments and hooks. For SDK authors, and for anyone who wants to know exactly how a flag is decided.