Core Concepts

Flagship is an AI-powered release engineering platform that transforms how teams ship software. It combines standard feature flagging with intelligent automation to ensure every deployment is safe, observable, and reversible.

The Flagship Workflow

Unlike traditional feature flag management where flags are manually created and managed, Flagship automates the lifecycle:

1. Detection: Flagship scans your Pull Requests to detect new code paths and automatically suggests feature flag wrapping.
2. Planning: AI analyzes the risk of the change and generates a rollout plan (e.g., "1% canary, then 10%, then 100%").
3. Rollout: The platform executes the plan automatically, monitoring metrics at each stage.
4. Safety: If anomalies or errors spike, Flagship leverages "Automated Rollback Triggers" to instantly revert the change.

Key Entities

Feature Flags

The fundamental unit of control. A feature flag decouples deployment (shipping code) from release (enabling features). Flagship flags are built on the OpenFeature standard, ensuring zero vendor lock-in.

Rollouts

A Rollout is the active process of releasing a feature. It is not just "on" or "off"; it is a state machine that moves through stages (Planning, Active, Paused, Completed).

Dimensions & Targeting

Rules that determine who sees a feature.

• Dimensions: Attributes of your user or request (e.g., user_id, email_domain, subscription_plan).
• Target Groups: Reusable sets of rules (e.g., "Internal Testers", "Beta Users").

Experiments (Coming Soon)

Advanced analysis that compares the performance of flagged vs. unflagged users to prove business impact.