Documentation Overview¶
Start with the task in front of you. Each path links to the guides that explain the relevant APIs, runtime behavior, and trade-offs.
Model a domain¶
Use the domain libraries for behavior, invariants, persistence boundaries, and business decisions.
- Domain covers aggregates, entities, value objects, and typed identifiers.
- Rules and Domain Policies model decisions and invariant checks.
- Domain Specifications encapsulate predicates and query intent.
- Domain Repositories and ActiveEntity provide two persistence models.
- Domain Events and Domain Change History capture state transitions.
Handle requests and workflows¶
Use the application libraries to dispatch work, return explicit outcomes, and coordinate multi-step processes.
- Results defines the success and failure model used across the kit.
- Application Commands and Queries defines request contracts and handlers.
- Requester and Notifier dispatches requests and notifications through pipeline behaviors.
- Pipelines coordinates observable in-process steps.
- Orchestrations handles durable workflows with signals, timers, and explicit state.
Integrate storage and messaging¶
Choose an integration by the data shape and delivery behavior your application needs.
- Messaging delivers messages to subscribers. Queueing assigns work to one consumer.
- Blob Storage stores named binary content. Document Storage stores documents and metadata.
- FileStorage provides file-oriented providers and operations.
- Jobs and Job Scheduling run background work now or on a schedule.
- Modules composes feature registrations and keeps module boundaries explicit.
Operate and diagnose an application¶
Use the host and tooling features to expose endpoints, inspect runtime state, and find failures.
- Presentation configures the host and its feature builders.
- Developer Dashboards exposes local pages for health, logs, metrics, jobs, queues, and other registered features.
- Log Entries, Metrics, and Correlation IDs connect requests to runtime evidence.
- Startup Tasks runs ordered work during application startup.
- CLI and Console Commands provide local operational commands.
Work with AI agents¶
The bdk mcp command connects MCP-capable coding agents to official DevKit documentation and local runtime diagnostics.
Start with AI Agent Support for the possibilities, installation, host registration, client configuration, and first self-test. Continue with the AI Agent Support Guide for the tool catalog, coding workflow, prompt library, project-owned diagnostics, and safety model.
Use the MCP Reference for the command and tool contract. Use MCP Client Configuration for supported client files.
Browse the complete reference¶
- Open the Complete Feature Index when you know the capability name.
- Open the API Reference when you need a namespace, type, or member signature.
- Use site search to find a symbol or phrase across both the authored pages and the feature guides.