Cratis Chronicle

Cross-references every registered event type against all projections, reducers, and reactors, and reports the event types that nothing consumes (data being written that no read model or reactor reads) plus the inverse: consumers that reference an event type id which no longer exists. Use as a periodic data-debt health check. Read-only introspection — grounded entirely in the store's registry and observer definitions.

Describes a registered event type's real schema — every property with its JSON type and a suggested C# type — read from the store's event type registry. Use this to ground any generated projection, read model, or spec in the event's actual fields instead of guessing. Returns null when the event type is not registered.

Deduces what the system in an event store is and is for. Clusters event types into entities by name, places every event in its entity's lifecycle (creation, mutation, activity, correction, termination), and maps read surfaces and automations. Use it to describe a system in domain language or to tell the story of how the system behaves — the result carries narrative guidance for doing exactly that. Read-only introspection.

Reads the ordered event history for one event source together with each event's correlation id and causation chain, so an agent can narrate what happened and why instead of a support engineer reading raw JSON logs. Read-only. Follow correlation ids to relate an event to the wider flow that produced it.

Produces a living data dictionary from the event store: every event type with its fields (from the registered schema) and every projection, reducer, and reactor that consumes it. Ideal onboarding material and grounding for modeling questions. Read-only introspection. Optionally scope to a single event type.

Retrieves events from an event sequence with optional filtering by sequence range, event source id, and event type. Returns event headers and JSON content.

Gets a specific job by ID. Returns either a job descriptor with full details including status changes and progress, or a job error (NotFound).

Gets the job steps for a specific job. Returns a collection of job step descriptors with status, progress, and metadata. Optionally filtered by step status.

Shows detailed information about a specific observer including its type, running state, owner, handled sequence numbers and observed event types.

Gets learned behavior patterns matching a grouping key and context, in the server's authoritative order. Returns full pattern details; an action is an ordinary facet, not a separate field.

Lists the current instances of a read model as JSON, with pagination. Use to inspect the projected state of entities. Fails for client-owned read models (see 'queryable' in list_read_models).

Gets version information from the connected Chronicle server. Use to verify connectivity and check the server version and commit.

Returns the highest used sequence number (tail) in an event sequence. This is not a total count of events — gaps may exist. Use to gauge how far a sequence has progressed.

Lists all event stores registered on the Chronicle server. Use to discover valid event store names.

Lists all registered event types in an event store, including their id, generation, owner and source. Use to explore the domain schema.

Lists observer partitions that have failed and are paused, including the failure reason and the sequence number where the failure occurred. Use to discover what needs investigation or recovery.

Lists all jobs in a namespace, optionally filtered by status. Returns a collection of job descriptors with status, progress, and metadata.

Lists all namespaces within an event store. Use to discover valid namespace names.

Lists observers (projections, reactors, reducers, client observers) in a namespace, with health and replay status. Optionally filter by type.

Lists all projection definitions registered in an event store, including the read model they project into and their activity state. Use to audit registered projections.

Lists all read model definitions registered in an event store. The 'queryable' field is false for client-owned read models whose state is stored by the client application rather than the Chronicle server; querying their instances will fail.

Lists active recommendations from the Chronicle server. Recommendations are automated suggestions for maintenance tasks such as rerunning projections or handling schema migrations.

Answers ad-hoc queries by projecting events on demand: reads the given event types from the store and folds them into per-event-source instances with AutoMap semantics (later events overwrite earlier property values; removedWith event types drop the instance). Nothing is registered or written. First resolve which event types are involved (describe_system or generate_event_catalog), then call this with them. Use it for natural-language questions like 'show all employees with all details'.

Generates a reviewable read model + model-bound projection from one or more event types, grounded in the events' real schema (field names and types come from the store, never guesses). Resolve the event types from the user's request first — with list_event_types, describe_event_type, or generate_event_catalog — then pass them here. Surfaces existing projections that already read these events so a duplicate is not created. When none of the requested event types are registered it returns no code and says so rather than fabricating a shape.

Suggests event types to introduce next for the system in an event store. Deduces entities and lifecycles from the registered event type names and schemas, then finds concrete gaps: entities without an explicit creation event, lifecycles with no termination, and creation-time properties no event can ever change. Each suggestion carries the gap it closes and candidate properties from the existing schema. Read-only introspection; refine the results with domain knowledge.