LaunchDarkly SDK Feature Comparison

AI Completion Configuration Send prompts to an LLM to generate a single response. Application metadata Specify application and application version information. Automatic environment attributes Automatically include device and application data in each evaluated context. Big segments Configure a persistent store to hold segments that are either synced from external tools, or that contain an arbitrarily large number of contexts of any one context kind. Bootstrapping Provide an initial set of flag values that are immediately available during client initialization. Contexts and context kinds Evaluate flags based on contexts. A context is a generalized way of referring to the people, services, machines, or other resources that encounter feature flags. SDKs without this feature only support flag evaluation for users. Data Saving Mode Our second generation flag delivery protocol. Event Compression Optionally support compressing event payloads sent to LaunchDarkly. Experimentation Connect a flag with one or more metrics to measure end-user behavior for different variations of a flag. Requires minimum SDK versions, but no SDK configuration. Flag evaluation reasons Receive information about how a flag variation was calculated, for example, because it matched a specific targeting rule. Generate bootstrap details Return the flag variations for all feature flags for a given context. Hooks Define callbacks that are executed by the SDK at various points of interest, usually for metrics or tracing. Inline Context In Custom Events The SDK will inline the context inside custom events. Inline Context In Evaluation Events The SDK will inline the context inside evaluation events. Migration flags Configure how to read and write data for an old and new system, determine which stage of a migration the application is in, execute the correct read and write calls for each stage. Multiple environments Evaluate flags from multiple environments using a single client instance Offline mode Close the SDK's connection to LaunchDarkly. Use cached or fallback values for each flag evaluation. Omit Anonymous Contexts from Events Optionally omit anonymous contexts from events. OpenTelemetry Add flag evaluation information to OpenTelemetry spans. Plugin Support Provides SDK extensibility through the use of plugins. Polling Response Compression Supports receiving compressed polling responses. Private attributes Use context attribute values for targeting, but do not send them to LaunchDarkly. Reading flags from a file Use flag values, specified in JSON or YAML files, for all flag evaluations. Useful for testing or prototyping; do not use in production. Relay Proxy in daemon mode Configure the SDK to connect to a persistent store, such as the Relay Proxy's data store. Relay Proxy in proxy mode Configure the SDK to connect to alternate service endpoints, such as the Relay Proxy. Secure mode For clent-side SDKs, require a hash, signed with the SDK key for the LaunchDarkly environment, to evaluate flag variations. For server-side or edge SDKs, generate a secure mode hash. Sending custom events Record actions taken in your application as events. You can connect to these events to metrics for use in experiments. Storing data Configure an external database as a feature store. Persist flag data across application restarts. Storing data (Consul) Configure Consul as a feature store. Persist flag data across application restarts. Storing data (DynamoDB) Configure DynamoDB as a feature store. Persist flag data across application restarts. Storing data (Redis) Configure Redis as a feature store. Persist flag data across application restarts. Subscribing to flag changes Use a listener pattern to subscribe to flag change notifications. Test data sources Mock data of an SDK. Useful for unit tests; cannot be used in production. Tracking AI metrics Collect and report metrics specific to AI/ML workloads and flag evaluations. Web proxy configuration Configure the SDK to connect to LaunchDarkly through a web proxy.

AI Completion Configuration Send prompts to an LLM to generate a single response. Application metadata Specify application and application version information. Automatic environment attributes Automatically include device and application data in each evaluated context. Big segments Configure a persistent store to hold segments that are either synced from external tools, or that contain an arbitrarily large number of contexts of any one context kind. Bootstrapping Provide an initial set of flag values that are immediately available during client initialization. Contexts and context kinds Evaluate flags based on contexts. A context is a generalized way of referring to the people, services, machines, or other resources that encounter feature flags. SDKs without this feature only support flag evaluation for users. Data Saving Mode Our second generation flag delivery protocol. Event Compression Optionally support compressing event payloads sent to LaunchDarkly. Experimentation Connect a flag with one or more metrics to measure end-user behavior for different variations of a flag. Requires minimum SDK versions, but no SDK configuration. Flag evaluation reasons Receive information about how a flag variation was calculated, for example, because it matched a specific targeting rule. Generate bootstrap details Return the flag variations for all feature flags for a given context. Hooks Define callbacks that are executed by the SDK at various points of interest, usually for metrics or tracing. Inline Context In Custom Events The SDK will inline the context inside custom events. Inline Context In Evaluation Events The SDK will inline the context inside evaluation events. Migration flags Configure how to read and write data for an old and new system, determine which stage of a migration the application is in, execute the correct read and write calls for each stage. Multiple environments Evaluate flags from multiple environments using a single client instance Offline mode Close the SDK's connection to LaunchDarkly. Use cached or fallback values for each flag evaluation. Omit Anonymous Contexts from Events Optionally omit anonymous contexts from events. OpenTelemetry Add flag evaluation information to OpenTelemetry spans. Plugin Support Provides SDK extensibility through the use of plugins. Polling Response Compression Supports receiving compressed polling responses. Private attributes Use context attribute values for targeting, but do not send them to LaunchDarkly. Reading flags from a file Use flag values, specified in JSON or YAML files, for all flag evaluations. Useful for testing or prototyping; do not use in production. Relay Proxy in daemon mode Configure the SDK to connect to a persistent store, such as the Relay Proxy's data store. Relay Proxy in proxy mode Configure the SDK to connect to alternate service endpoints, such as the Relay Proxy. Secure mode For clent-side SDKs, require a hash, signed with the SDK key for the LaunchDarkly environment, to evaluate flag variations. For server-side or edge SDKs, generate a secure mode hash. Sending custom events Record actions taken in your application as events. You can connect to these events to metrics for use in experiments. Storing data Configure an external database as a feature store. Persist flag data across application restarts. Storing data (Consul) Configure Consul as a feature store. Persist flag data across application restarts. Storing data (DynamoDB) Configure DynamoDB as a feature store. Persist flag data across application restarts. Storing data (Redis) Configure Redis as a feature store. Persist flag data across application restarts. Subscribing to flag changes Use a listener pattern to subscribe to flag change notifications. Test data sources Mock data of an SDK. Useful for unit tests; cannot be used in production. Tracking AI metrics Collect and report metrics specific to AI/ML workloads and flag evaluations. Web proxy configuration Configure the SDK to connect to LaunchDarkly through a web proxy.