使用说明

SF Data 360 MCP Server

An MCP server that lets AI agents and engineers inspect, query, and operate Salesforce Data 360 through authenticated, bounded workflows.

It turns a large enterprise API surface into repeatable operator actions: discover schemas, validate queries, inspect identity and segmentation configuration, manage activations and pipelines, and troubleshoot live environments without building a Connected App.

Why it exists

Enterprise data work is rarely one API call. Operators need to select the right org, understand the data model, validate inputs, keep outputs manageable, and preserve a clear boundary around consequential changes. This server packages those steps as agent-ready tools using existing Salesforce CLI authentication.

What it enables

Grounded discovery — Inspect schemas, relationships, metadata, and sample data before acting.
Repeatable operations — Use consistent workflows across ingestion, harmonization, identity, segmentation, activation, AI/ML, and administration.
Multi-org delivery — Switch explicitly between authenticated environments.
Safer agent use — Apply bounded outputs, validation, transient-failure retries, and explicit target-org selection.
Full lifecycle coverage — Operate the broader Connect API lifecycle without making command counts the product story.

Quick Start

Prerequisites

Python 3.10+
Salesforce CLI installed
A Salesforce org with Data Cloud enabled

Installation

git clone https://github.com/stijnhoste/sf-data-360-mcp.git
cd sf-data-360-mcp
pip install -r requirements.txt

# Authenticate with your Data Cloud org
sf org login web --alias my-dc-org

Configure Your MCP Client

Add to your MCP client configuration file. You can use either:

Option 1: Project-level config (recommended for team projects)

.mcp.json              # In your project root

Option 2: Global config (applies to all projects)

Client	Config File
Cursor	`~/.cursor/mcp.json`
Claude Code	`~/.claude/mcp.json`

Add this configuration:

{
  "mcpServers": {
    "datacloud": {
      "command": "python",
      "args": ["/absolute/path/to/sf-data-360-mcp/server.py"],
      "env": {
        "DC_DEFAULT_ORG": "my-dc-org"
      }
    }
  }
}

Note: Replace /absolute/path/to/ with the actual path where you cloned the repository. Project-level .mcp.json takes precedence over global config when both exist.

Configuration

Environment Variables

Variable	Required	Default	Description
`DC_DEFAULT_ORG`	No	-	SF CLI org alias to use by default
`DC_API_VERSION`	No	`v65.0`	Salesforce API version for Connect API
`DC_RETRY_TOTAL`	No	`3`	Number of retries for transient failures (5xx)
`DC_RETRY_BACKOFF`	No	`0.5`	Backoff factor between retries (seconds)
`DC_RATE_LIMIT_REQUESTS`	No	`100`	Max requests per rate limit window
`DC_RATE_LIMIT_WINDOW`	No	`20`	Rate limit window in seconds
`DEFAULT_LIST_TABLE_FILTER`	No	`%`	SQL LIKE pattern for filtering tables

Multi-Org Support

If you don't set DC_DEFAULT_ORG, use these tools to select an org at runtime:

list_orgs()           # See all authenticated orgs
set_target_org(alias) # Switch to a specific org
get_target_org()      # See currently selected org

API Capabilities

This MCP server uses the Connect API exclusively (works with SF CLI auth) and now supports full CRUD operations for most resources:

Operation	Supported	Notes
Read (list, get, query)	✅ Yes	All resources
Create	✅ Yes	Most resources via Connect API
Update	✅ Yes	All resources
Delete	✅ Yes	All resources
Run/Execute	✅ Yes	Transforms, insights, identity resolution

Note: Some complex resources may still require Metadata API or Setup UI for initial creation, but ongoing management is fully supported.

Tool reference

Org Management (3)

Tool	Description
`list_orgs()`	List all SF CLI authenticated orgs
`set_target_org(alias)`	Switch to a different org
`get_target_org()`	Get currently selected org

Query Tools (6)

Tool	Description
`query(sql)`	Execute SQL queries against Data Cloud
`query_v2(sql)`	Execute SQL using V2 query API
`get_query_batch_v2(query_id)`	Get results from V2 query batch
`cancel_sql_query(query_id)`	Cancel a running SQL query
`validate_query(sql)`	Validate SQL syntax before execution
`format_sql(sql)`	Format SQL for readability

Schema Discovery (8)

Tool	Description
`list_tables()`	List available tables
`describe_table(table)`	Get column names
`describe_table_full(table)`	Get detailed schema with types
`get_metadata(...)`	Get rich entity metadata
`get_relationships(entity)`	Get relationships for JOINs
`get_dmo_relationships(dmo)`	Get DMO relationships
`explore_table(table)`	Schema + samples + profiles
`search_tables(keyword)`	Search tables/columns

Segments (9)

Tool	Description
`list_segments()`	List all segments
`get_segment(name)`	Get segment details
`get_segment_members(name)`	Get segment members
`count_segment(name)`	Count segment members
`create_segment(definition)`	Create a new segment
`update_segment(name, updates)`	Update segment
`delete_segment(name)`	Delete segment
`deactivate_segment(name)`	Deactivate a segment
`publish_segment(name)`	Publish a segment

Activations (11)

Tool	Description
`list_activations()`	List all activations
`get_activation(id)`	Get activation details
`create_activation(definition)`	Create a new activation
`update_activation(id, updates)`	Update activation
`delete_activation(id)`	Delete activation
`get_activation_data(id)`	Get activation data
`list_activation_targets()`	List activation targets
`get_activation_target(name)`	Get activation target details
`create_activation_target(definition)`	Create activation target
`update_activation_target(name, updates)`	Update activation target
`list_activation_external_platforms()`	List external platforms

Data Streams (6)

Tool	Description
`list_data_streams()`	List all data streams
`get_data_stream(name)`	Get stream details
`create_data_stream(definition)`	Create a new data stream
`update_data_stream(name, updates)`	Update data stream
`delete_data_stream(name)`	Delete data stream
`run_data_stream(name)`	Run a data stream

Data Transforms (13)

Tool	Description
`list_data_transforms()`	List all transforms
`get_data_transform(name)`	Get transform details
`create_data_transform(definition)`	Create a new transform
`update_data_transform(name, updates)`	Update transform
`delete_data_transform(name)`	Delete transform
`run_data_transform(name)`	Run transform
`cancel_data_transform(name)`	Cancel running transform
`retry_data_transform(name)`	Retry failed transform
`refresh_data_transform_status(name)`	Refresh transform status
`get_transform_run_history(name)`	Get run history
`get_transform_schedule(name)`	Get transform schedule
`update_transform_schedule(name, schedule)`	Update schedule
`validate_data_transform(definition)`	Validate transform definition

Connections (11)

Tool	Description
`list_connections()`	List all connections
`get_connection(name)`	Get connection details
`get_connection_objects(name)`	Get available objects
`get_connection_schema(name, object)`	Get object schema
`get_connection_databases(name)`	Get available databases
`get_connection_database_schemas(name, db)`	Get database schemas
`get_connection_endpoints(name)`	Get connection endpoints
`list_connectors()`	List connector types
`get_connector(type)`	Get connector details
`update_connection(name, updates)`	Update connection
`delete_connection(name)`	Delete connection

Data Lake Objects (5)

Tool	Description
`list_data_lake_objects()`	List all DLOs
`get_data_lake_object(name)`	Get DLO details
`create_data_lake_object(definition)`	Create a new DLO
`update_data_lake_object(name, updates)`	Update DLO
`delete_data_lake_object(name)`	Delete DLO

Data Model Objects (14)

Tool	Description
`list_data_model_objects()`	List all DMOs
`get_data_model_object(name)`	Get DMO details
`create_data_model_object(definition)`	Create a new DMO
`update_data_model_object(name, updates)`	Update DMO
`delete_data_model_object(name)`	Delete DMO
`get_dmo_relationships(name)`	Get DMO relationships
`create_dmo_relationship(dmo, definition)`	Create DMO relationship
`delete_dmo_relationship(name)`	Delete DMO relationship
`list_data_model_object_mappings()`	List DMO mappings
`get_data_model_object_mapping(name)`	Get DMO mapping
`create_data_model_object_mapping(definition)`	Create DMO mapping
`delete_dmo_mapping(name)`	Delete DMO mapping
`delete_data_model_field_mappings(name)`	Delete field mappings
`update_data_model_field_mappings(name, mappings)`	Update field mappings

Data Spaces (7)

Tool	Description
`list_data_spaces()`	List all data spaces
`get_data_space(name)`	Get data space details
`create_data_space(definition)`	Create a new data space
`update_data_space(name, updates)`	Update data space
`get_data_space_members(name)`	Get objects in space
`get_data_space_member(space, member)`	Get specific member
`update_data_space_members(name, members)`	Update space members

Calculated Insights (8)

Tool	Description
`list_calculated_insights()`	List all insights
`get_calculated_insight(name)`	Get insight details
`create_calculated_insight(definition)`	Create a new insight
`update_calculated_insight(name, updates)`	Update insight
`delete_calculated_insight(name)`	Delete insight
`run_calculated_insight(name)`	Run insight refresh
`query_calculated_insight(...)`	Query insight data
`get_insight_metadata(name)`	Get insight metadata

Data Graphs (7)

Tool	Description
`list_data_graphs()`	List all data graphs
`get_data_graph(name)`	Get data graph details
`get_data_graph_metadata()`	Get data graph metadata
`create_data_graph(definition)`	Create a new data graph
`delete_data_graph(name)`	Delete data graph
`refresh_data_graph(name)`	Refresh data graph
`query_data_graph(...)`	Query unified profiles

Identity Resolution (7)

Tool	Description
`list_identity_rulesets()`	List identity rulesets
`get_identity_ruleset(name)`	Get ruleset details
`create_identity_ruleset(definition)`	Create a new ruleset
`update_identity_ruleset(name, updates)`	Update ruleset
`delete_identity_ruleset(name)`	Delete ruleset
`run_identity_resolution(name)`	Run identity resolution
`lookup_unified_id(...)`	Look up unified IDs

Profile Queries (4)

Tool	Description
`query_profile(dmo, limit, offset)`	Query profile records
`get_profile_record(dmo, id)`	Get specific profile record
`get_profile_record_with_children(dmo, id, child)`	Get profile with children
`get_profile_record_with_insights(dmo, id, insight)`	Get profile with insights

AI & ML - Models (6)

Tool	Description
`list_ml_models()`	List all ML models
`get_ml_model(name)`	Get model details
`get_prediction(model, data)`	Get predictions
`update_ml_model(name, updates)`	Update ML model
`delete_ml_model(name)`	Delete ML model
`list_model_artifacts()`	List model artifacts
`get_model_artifact(name)`	Get artifact details
`update_model_artifact(name, updates)`	Update artifact
`delete_model_artifact(name)`	Delete artifact

AI & ML - Document AI (6)

Tool	Description
`list_document_ai_configs()`	List Document AI configs
`get_document_ai_config(name)`	Get config details
`get_document_ai_global_config()`	Get global Document AI config
`extract_document_data(...)`	Extract document data
`run_document_ai(config, data)`	Run Document AI extraction
`generate_document_schema(config)`	Generate document schema
`update_document_ai_config(name, updates)`	Update config
`delete_document_ai_config(name)`	Delete config

AI & ML - Semantic Search (5)

Tool	Description
`list_semantic_searches()`	List semantic search indexes
`get_semantic_search(name)`	Get search index details
`get_semantic_search_config()`	Get global search config
`update_semantic_search(name, updates)`	Update search index
`delete_semantic_search(name)`	Delete search index

Data Actions (11)

Tool	Description
`list_data_actions()`	List event-driven automation rules
`get_data_action(name)`	Get data action details
`create_data_action(definition)`	Create a new data action
`update_data_action(name, updates)`	Update data action
`delete_data_action(name)`	Delete data action
`list_data_action_targets()`	List webhook/external targets
`get_data_action_target(name)`	Get target details
`create_data_action_target(definition)`	Create a new target
`delete_data_action_target(name)`	Delete target
`get_data_action_target_signing_key(name)`	Get target signing key

Admin & Monitoring (9)

Tool	Description
`get_limits()`	Get API limits and usage
`get_data_kit_component_dependencies(kit, component, type)`	Get data kit dependencies
`get_data_kit_deployment_status(kit, component, type)`	Get deployment status
`undeploy_data_kit_component(kit, definition)`	Undeploy data kit components
`list_private_network_routes()`	List private network routes
`get_private_network_route(id)`	Get route details
`create_private_network_route(definition)`	Create a network route
`delete_private_network_route(id)`	Delete a network route

Data Clean Rooms (19)

Tool	Description
`list_data_clean_room_collaborations()`	List all collaborations
`get_data_clean_room_collaboration(id)`	Get collaboration details
`create_data_clean_room_collaboration(definition)`	Create a collaboration
`update_data_clean_room_collaboration(id, updates)`	Update collaboration
`delete_data_clean_room_collaboration(id)`	Delete collaboration
`accept_data_clean_room_collaboration_invitation(id)`	Accept invitation
`reject_data_clean_room_collaboration_invitation(id)`	Reject invitation
`run_data_clean_room_collaboration(id)`	Run collaboration
`list_data_clean_room_collaboration_jobs(id)`	List collaboration jobs
`get_data_clean_room_query_job(collaboration, job)`	Get query job
`list_data_clean_room_providers()`	List all providers
`get_data_clean_room_provider(id)`	Get provider details
`create_data_clean_room_provider(definition)`	Create a provider
`update_data_clean_room_provider(id, updates)`	Update provider
`delete_data_clean_room_provider(id)`	Delete provider
`list_data_clean_room_provider_templates(id)`	List provider templates
`list_data_clean_room_templates()`	List all templates
`list_data_clean_room_specifications()`	List specifications
`create_data_clean_room_specification(definition)`	Create specification

Raw Connect API (1)

Tool	Description
`connect_api_request(method, endpoint, params, body)`	Make raw Connect API request

autoApprove Settings

For agentic workflows, auto-approve read-only tools:

{
  "autoApprove": [
    "list_orgs", "get_target_org",
    "list_tables", "describe_table", "describe_table_full",
    "get_metadata", "get_relationships", "get_dmo_relationships",
    "explore_table", "search_tables",
    "list_segments", "get_segment", "get_segment_members", "count_segment",
    "list_activations", "get_activation", "get_activation_data",
    "list_activation_targets", "get_activation_target", "list_activation_external_platforms",
    "list_data_streams", "get_data_stream",
    "list_data_transforms", "get_data_transform", "get_transform_run_history", "get_transform_schedule",
    "list_connections", "get_connection", "get_connection_objects", "get_connection_schema",
    "get_connection_databases", "get_connection_database_schemas", "get_connection_endpoints",
    "list_connectors", "get_connector",
    "list_data_lake_objects", "get_data_lake_object",
    "list_data_model_objects", "get_data_model_object", "get_dmo_relationships",
    "list_data_model_object_mappings", "get_data_model_object_mapping",
    "list_data_spaces", "get_data_space", "get_data_space_members", "get_data_space_member",
    "list_calculated_insights", "get_calculated_insight", "get_insight_metadata",
    "list_data_graphs", "get_data_graph", "get_data_graph_metadata",
    "list_identity_rulesets", "get_identity_ruleset",
    "query_profile", "get_profile_record", "get_profile_record_with_children", "get_profile_record_with_insights",
    "list_ml_models", "get_ml_model", "list_model_artifacts", "get_model_artifact",
    "list_document_ai_configs", "get_document_ai_config", "get_document_ai_global_config",
    "list_semantic_searches", "get_semantic_search", "get_semantic_search_config",
    "list_data_actions", "get_data_action",
    "list_data_action_targets", "get_data_action_target", "get_data_action_target_signing_key",
    "list_private_network_routes", "get_private_network_route",
    "list_data_clean_room_collaborations", "get_data_clean_room_collaboration",
    "list_data_clean_room_collaboration_jobs", "get_data_clean_room_query_job",
    "list_data_clean_room_providers", "get_data_clean_room_provider",
    "list_data_clean_room_provider_templates", "list_data_clean_room_templates",
    "list_data_clean_room_specifications",
    "get_limits", "get_data_kit_component_dependencies", "get_data_kit_deployment_status",
    "validate_query", "format_sql", "validate_data_transform"
  ]
}

Authentication

This project uses SF CLI authentication exclusively:

Authenticate via SF CLI: sf org login web --alias my-org
Set DC_DEFAULT_ORG environment variable, or use set_target_org() tool
The server reads credentials from SF CLI's secure token storage

Benefits:

No Connected App setup required
No client secrets to manage
Multi-org support built-in
Tokens managed by SF CLI

API Architecture

All tools use the Connect API (/services/data/v65.0/ssot/*), which works with standard Salesforce OAuth tokens from SF CLI.

Connect API (/services/data/v65.0/ssot/*)
├── query-sql/*              → query()
├── metadata                 → get_metadata(), describe_table_full()
│
├── segments/*               → list_segments(), create_segment(), etc.
├── activations/*            → list_activations(), create_activation(), etc.
│
├── data-streams/*           → list_data_streams(), create_data_stream()
├── data-transforms/*        → list_data_transforms(), create_data_transform()
├── connections/*            → list_connections(), get_connection()
│
├── data-lake-objects/*      → list_data_lake_objects(), create_data_lake_object()
├── data-model-objects/*     → list_data_model_objects(), create_data_model_object()
├── data-spaces/*            → list_data_spaces(), create_data_space()
│
├── calculated-insights/*    → list_calculated_insights(), create_calculated_insight()
├── data-graphs/*            → list_data_graphs(), create_data_graph()
├── identity-resolutions/*   → list_identity_rulesets(), create_identity_ruleset()
├── universalIdLookup/*      → lookup_unified_id()
│
├── machine-learning/*       → list_ml_models(), get_prediction()        ─┐
├── document-processing/*    → list_document_ai_configs()                 │ AI & ML
├── search-index/*           → list_semantic_searches()                  ─┘
│
├── data-actions/*           → list_data_actions(), create_data_action() ─┐
├── data-action-targets/*    → list_data_action_targets()                 │ Data Actions
├── private-network-routes/* → list_private_network_routes()             ─┘
│
├── data-clean-room/*        → list_data_clean_room_collaborations()     ─ Clean Rooms
│
└── /services/data/v65.0/limits → get_limits()                           ─ Admin

Contributing

See CONTRIBUTING.md for guidelines.

License

MIT License