FeatBit review: open-source, self-hosted feature flag platform with A/B testing, targeting, and AI-native release tools. See pricing, features, and…
FeatBit is an open-source feature flag and experimentation platform that lets engineering teams separate code deployment from feature release. Teams wrap new functionality behind a flag, then use FeatBit's console or API to decide who sees it, when, and at what rollout percentage, without needing a new deployment to change behavior in production.
The project is built primarily in C# on .NET with an Angular-based web console, and it stores data in MongoDB (with documented support for Azure Cosmos DB and PostgreSQL via FerretDB), using Redis for caching and messaging. Most of the codebase is released under the MIT license, and the platform can be run entirely self-hosted via Docker Compose, Kubernetes/Helm, Terraform on AWS, or Azure Container Apps, or consumed as a managed cloud service.
A defining technical trait is real-time flag propagation over WebSocket connections rather than SDK polling, with the company stating that flag changes reach connected SDKs in under 100 milliseconds. This is combined with SDKs for .NET, JavaScript, React, React Native, Node.js, Java, Python, and Go, plus OpenFeature provider support for teams standardizing on the vendor-neutral OpenFeature spec.
Beyond basic on/off toggles, FeatBit supports progressive rollouts starting at as little as one percent of traffic, individual user targeting, reusable audience segments, multivariate flags for A/B and multivariate experiments, and built-in experimentation tooling for measuring the impact of a release against product metrics.
Governance features include audit logs, scheduled flag changes, and change-approval workflows, alongside integrations with Datadog, New Relic, Grafana, and Slack, and OpenTelemetry support for tracing. A Relay Proxy ('FeatBit Agent') caches flag data closer to application servers for distributed or edge deployments.
FeatBit has more recently added AI-native tooling, including a hosted Model Context Protocol (MCP) server at mcp.featbit.co that lets AI coding agents like Claude Code create, evaluate, list, and update feature flags via natural-language instructions, framed around use cases such as controlling which versions of AI agent skills or tools are active for different user cohorts.
FeatBit's self-hosted pricing is structured around feature access and support level rather than usage-based metering: the company states that no self-hosted plan charges based on seats, monthly active users, or API calls.
The free Foundation tier is the full open-source, MIT-licensed platform, self-hosted with complete data ownership, including the core console, REST API, SDKs, webhooks, real-time WebSocket updates, basic role-based access control, and community support.
The Enterprise tier is priced at 3,999 dollars per year and adds email and priority support, audit logs, flag approval workflows, a high-availability advisor, and up to 99 auto-sync Relay Proxy agents, with advanced role-based access control available as a roughly 600-dollar-per-year add-on. An Enterprise Premium tier offers custom pricing for large-scale deployments needing 100-plus auto-sync agents, dedicated engineering support, and custom SLAs. A cloud-hosted (SaaS) option with annual discounts is also available by contacting FeatBit's sales team.
Yes. FeatBit's Foundation tier is the full open-source platform released under the MIT license, and it is free to self-host indefinitely. Paid Enterprise and Enterprise Premium tiers add support, governance, and scaling features on top of the same core platform.
Yes, self-hosting is central to FeatBit's design. It can be deployed via Docker Compose, Kubernetes using Helm charts, Terraform modules for AWS, or Azure Container Apps, giving teams full control over where flag data and evaluation logic run.
FeatBit provides official SDKs for .NET/C#, JavaScript, React, React Native, Node.js, Java, Python, and Go, and it also supports OpenFeature providers for teams using the vendor-neutral OpenFeature specification.
Unlike many commercial feature flag vendors, FeatBit states that it does not charge based on seats, monthly active users, or API calls. Its self-hosted Enterprise tier is a flat annual fee (3,999 dollars per year at time of research), which the company positions as significantly cheaper at scale than seat- or MAU-metered competitors like LaunchDarkly.
Yes. FeatBit includes multivariate flags and built-in experimentation tooling that let teams run A/B tests and measure the impact of a rollout against product metrics, in addition to standard percentage-based progressive rollouts.
Yes. FeatBit publishes a hosted Model Context Protocol (MCP) server that lets AI coding agents such as Claude Code create, evaluate, list, and update feature flags through natural-language instructions, aligning with FeatBit's positioning around AI-native release engineering.