Bucketeer

Bucketeer is a self-hosted feature-flag and experimentation platform for teams that want control over how changes reach users. It suits engineering groups willing to operate their own release tooling and offers a broad set of rollout controls at no per-seat cost.

Overview

Teams can switch features on or off without deploying code, then use dark launches and staged rollouts to manage exposure. Targeting by user attributes, devices, and segments helps teams vary releases by audience rather than relying on a single global switch. The trade-off is operational responsibility: Docker Compose supports lightweight setups, while Kubernetes with Helm is the production deployment option. Bucketeer says the platform can scale to 100M+ users.

For A/B testing, Bucketeer uses Bayesian probabilities and says this approach can produce results with smaller sample sizes. That can be useful when traffic is limited, but teams still need to make their own experiment-design and interpretation decisions.

Key features

Rollouts, rules, and experiments

Progressive and percentage rollouts, reusable segments, date-based targeting, flag-based rules, and prerequisites provide a flexible release toolkit. Prerequisites can tie one flag to another, while staged exposure supports gradual releases. This range is useful for teams with audience-specific launches or dependent features; it may be more machinery than a team needs for occasional, simple toggles.

Automation and integrations

Auto operations can schedule flag changes, respond to event rates, and trigger webhooks for CI/CD or APM integrations such as Datadog, New Relic, and Dynatrace. Slack notifications, Firebase Cloud Messaging push updates, webhooks, a RESTful HTTP API, and OpenFeature compatibility extend the platform's connections to release workflows. The important caveat is that the HTTP API is labeled Alpha, so teams that depend on it should account for that status.

Access and evaluation

Environment-level Viewer, Editor, and Owner roles, audit logging, separate client, server, and public API keys, and environment isolation give teams useful governance controls. Server SDKs support local and remote evaluation, but client SDKs evaluate remotely. That distinction matters for architecture: teams that require local evaluation on client devices should look elsewhere.

SDKs cover Android, iOS, Flutter, JavaScript, React, React Native, Go, Node.js, and OpenFeature. Bucketeer directs questions and community discussion to its Cloud Native Slack channel.

Pricing

Bucketeer (open-source self-hosted) costs nothing, is free forever, and has no per-seat pricing. The plan includes open-source self-hosting, targeting rules, percentage rollouts, user segments, and an audit log. Its no-per-seat model suits teams that want to add collaborators without a seat charge, while self-hosting means they must run the platform themselves. No trial or renewal term applies to the free-forever plan.

Platforms

Bucketeer supports Android, iOS, web, API, and self-hosted deployments. Its client SDKs are for Android, iOS, Flutter, JavaScript, React, and React Native; server SDKs are for Go and Node.js. Client-side remote evaluation and server-side local or remote evaluation are the key practical differences between those SDK groups.

Who it's for

Bucketeer is a strong fit for engineering teams that can operate Docker Compose or Kubernetes deployments and want release controls, audience targeting, experimentation, and governance without per-seat fees. It is a weaker choice for teams that need client SDKs to evaluate locally or need to rely on an HTTP API with a more established status.

Pros and cons

• Pro: Progressive rollouts, segments, targeting rules, and prerequisites let teams control exposure in several ways.
• Pro: The free-forever, self-hosted plan has no per-seat pricing, which avoids a seat-based charge as the team grows.
• Pro: Role controls, audit logging, separate key types, and environment isolation support governed release workflows.
• Con: Teams must deploy and operate the service themselves, using Docker Compose or Kubernetes with Helm.
• Con: Client SDKs evaluate remotely, which rules out workflows that require client-side local evaluation.
• Con: The HTTP API is labeled Alpha, a meaningful caveat for API-dependent integrations.

Alternatives

Feature Flag Software and Experimentation Platforms are useful category lists for comparing options. Among the named alternatives, GrowthBook is worth considering for teams that want a self-hosted open-source plan with unlimited experiments and traffic in one project; its listed Starter cloud plan is also 0.00 USD per.

Confidence by Spotify may suit teams seeking a free hosted option with 75,000 credits per month, unlimited flags and seats, up to five surfaces, Google SSO, and a one-month audit log. Statsig is an alternative for teams that want a free developer plan with 2M events per month, unlimited flag and config checks, 50,000 session replays per month, one-year analytics retention, and unlimited seats.

Amplitude may fit teams looking for a free plan billed at 2M events per month, with 10,000 monthly session replays and 2,000 AI feedback records. Compose offers an Early Adopter Plan billed at $0.0012 per Monthly Tested User each month, charged after the month for prior usage, for teams considering usage-based experimentation pricing.

ABsmartly is a paid option with pricing by request and a yearly subscription. AB Tasty is another paid option with Pro and Growth plans at custom pricing. Marpipe is an alternative with free catalog creative and feed-management plans.

Verdict

Choose Bucketeer if your engineering team wants a free, self-hosted platform with flexible rollout controls, targeting, experimentation, and no per-seat charge. Look elsewhere if client-side local evaluation is essential or your workflow depends on an HTTP API beyond its Alpha stage.