MCP Server

Narrative Data Collaboration

Query datasets, run NQL, and orchestrate workflows on the Narrative platform from any MCP client.

About

Overview

The Narrative Data Collaboration MCP server exposes the Narrative platform to any MCP-compatible agent. Connect Claude Code, Claude Desktop, Cursor, or your own agent and let it search the dataset catalog, run and validate NQL, curate datasets and their mappings onto Rosetta Stone attributes, connect a dataset to a destination connector, train a classifier, and create, trigger and inspect workflows and their jobs — all against your own data, governed by your Narrative permissions.

Authentication is handled via OAuth against your Narrative account, so the agent only ever sees the data you can see.

MCPNQLDatasetsWorkflowsNarrative

MCP Server Support

Specifications

Transport: Streamable HTTP
Hosting: Narrative-managed
Authentication: OAuth
Tools: 38

Tools

Tool	Description
`narrative_access_rules_describe`	Describe access rules by id, including their exposed attributes.
`narrative_access_rules_search`	Search the access rules the company can query.
`narrative_app_interfaces_list`	List a connector app's interfaces and their quick-settings schemas.
`narrative_app_profiles_list`	List an installation's connector profiles.
`narrative_attribute_create`	Create a Rosetta Stone attribute owned by the active company.
`narrative_attributes_describe`	Describe attributes by id, including how they are derived.
`narrative_attributes_search`	Search the Rosetta Stone attribute catalog.
`narrative_classifier_train`	Train a classifier on a dataset and return the training job id.
`narrative_connection_create`	Connect a dataset to a connector app to route its data onward.
`narrative_context_get`	Get the session context: user, role, and active company.
`narrative_context_get_companies`	List the companies the authenticated user can act for.
`narrative_context_search_companies`	Search companies by name. Global admins only.
`narrative_context_set_company`	Switch the active company for this session.
`narrative_data_planes_describe`	Describe data planes by id, including their compute pools.
`narrative_data_planes_list`	List the data planes the company can reach.
`narrative_dataset_activate`	Activate a dataset, locking its schema so ingestion can start.
`narrative_dataset_get_column_stats`	Read per-column statistics for a dataset.
`narrative_dataset_get_compatible_interfaces`	Check which installed connector interfaces accept a dataset's schema.
`narrative_dataset_recalculate_statistics`	Enqueue a recalculation of a dataset's column statistics.
`narrative_dataset_refresh_materialized_view`	Refresh a materialized-view dataset.
`narrative_dataset_request_sample`	Enqueue a sampling job for a dataset.
`narrative_dataset_set_column_stats_config`	Replace a dataset's column-statistics configuration.
`narrative_dataset_update`	Update a dataset's display name, description, and tags.
`narrative_datasets_describe`	Describe a dataset's schema, statistics, and mappings.
`narrative_datasets_search`	Search and filter the dataset catalog.
`narrative_derivation_create`	Create a derivation rule. Global admins only.
`narrative_derivations_describe`	Describe derivation rules by id.
`narrative_derivations_list`	List the derivation rules the company can use.
`narrative_installations_list`	List the apps the company has installed.
`narrative_jobs_describe`	Describe platform jobs by id, including their results.
`narrative_jobs_search`	Search and filter platform jobs.
`narrative_mapping_create`	Map a dataset onto an attribute.
`narrative_nql_execute`	Submit an NQL query for execution and return the workflow that runs it.
`narrative_nql_validate`	Validate NQL without executing it.
`narrative_workflow_runs_list`	List runs of a workflow and their status.
`narrative_workflows_create`	Create a workflow from a YAML specification.
`narrative_workflows_describe`	Describe workflows by id.
`narrative_workflows_trigger`	Trigger a workflow run.