feat

feat is a hosted feature-flag service for early engineering teams that need fast server-side decisions and controlled releases. It suits developers who value broad SDK coverage, flexible rule logic, and assignments that remain stable across sessions. It is less compelling for organizations that require a fully developed experimentation suite or established compliance certifications today.

Overview

Backend SDKs evaluate flags locally, so applications do not make a network request for every read. Configuration changes stream to connected SDKs in under a second, combining low-latency checks with quick rollout updates. That architecture is practical for services that read flags frequently, although teams must still manage their own rollout logic and operational discipline.

Key features

Targeting and rollouts

Targeting rules combine conditions across users, devices, organizations, sessions, and requests with AND and OR operators. This supports precise, multi-context releases rather than only simple user attributes, but it places responsibility for designing and maintaining potentially complex rules on the engineering team.

Sticky percentage rollouts keep a user, device, or organization on the same variation across requests and sessions. Stable assignment helps staged launches remain diagnostically clear and avoids exposing one person to changing variants between visits.

SDKs and OpenFeature

First-party SDKs cover Node.js, Bun, Deno, Go, Python, Ruby, browser JavaScript, and Flutter/Dart. OpenFeature providers extend feat to the OpenFeature web, Node.js, and Dart SDKs, making it a reasonable fit for teams standardizing flag access across different runtimes. The separately stated SDK language list also includes JavaScript, Go, Python, Ruby, and Cloudflare Workers.

Integrations, privacy, and controls

Metric sources can be connected through Segment, Mixpanel, PostHog, or custom sources. A built-in statistical engine, variant tracking, metric integrations, and automatic advancement of gradual rollouts are planned for Q3 2026, so current users should not treat those as present experimentation capabilities. Slack integration is also planned for that quarter.

Backend SDKs send no evaluation context to feat. Browser SDKs report only hashed identifier-and-kind pairs for monthly active-user counting, reducing the data sent during client-side use. Customer-facing endpoints use TLS 1.2 or newer and HSTS, with managed-service encryption at rest for Postgres, R2, and KV. SOC 2 Type I scaffolding is in progress, Type II is planned after its observation window, and ISO 27001 remains on the roadmap; compliance-led buyers may therefore need a more mature provider.

Pricing

feat has a useful free entry point, with paid tiers scaling primarily by monthly active users (MAU), server connections, retention, and support.

Free is suitable for small products and prototypes, especially with unlimited environments and flags, but its 7,500-MAU and 100-connection ceilings and short audit retention limit larger operations. Starter adds capacity and email support; Growth is the more credible production tier when a 99.9% SLA and longer audit history matter. Scale and Elastic target high-traffic services, while Enterprise is for organizations needing identity, tenancy, or residency controls.

Platforms

feat supports Android, iOS, web, and API use. Its SDK ecosystem spans server, browser, edge, and Flutter/Dart applications, with Cloudflare Workers included in the stated language coverage.

Who it's for

Choose feat for an early-stage engineering organization that wants local server evaluation, rapid configuration propagation, multi-context targeting, and sticky releases without building those systems itself. It is a weaker choice for teams whose primary requirement is current, built-in experimentation analytics or completed formal certifications.

Pros and cons

Pros

• Local server evaluation avoids a network call on each flag read.
• Sub-second streaming updates keep connected SDKs current.
• Rules cover users, devices, organizations, sessions, and requests with AND/OR logic.
• Sticky rollouts preserve consistent assignments across requests and sessions.
• Free tier includes unlimited environments and flags.

Cons

• Experimentation and Slack capabilities are planned rather than current.
• SOC 2 and ISO 27001 work is not yet complete.
• Free limits of 7,500 MAU and 100 server connections can constrain growing services.
• Complex targeting is powerful but demands ongoing engineering ownership.

Alternatives

See the Feature Flag Software list for a broader comparison.

Flagsmith is a better fit when open-source functionality, self-hosting, or a wider platform matrix matters. Choose Featurevisor for a free, MIT-licensed self-hosted option with no per-seat pricing, or GO Feature Flag for open-source deployment on your own infrastructure with unlimited flags.

Unleash suits teams wanting an open-source plan and self-hosting choices. ConfigCat is worth considering for a forever-free plan centered on flags, environments, and configuration-download quotas. Featureflip offers a small free Solo tier, while Featureflow provides a free hosted plan with seats, projects, environments, and SDK-request limits. Flipt is an alternative for Git-native, self-hosted workflows.

Verdict

feat is a strong choice for early teams that want privacy-conscious local evaluation, flexible targeting, and dependable sticky rollouts across many SDKs. Look elsewhere if mature experimentation analytics or completed enterprise compliance certifications are immediate requirements.