salesforce-datacloud-mcp

MCP server for Salesforce Data Cloud (Data 360). Connects Claude Desktop (or any MCP client) to your Data Cloud org so you can explore and manage the full platform — data streams, DMOs, segments, calculated insights, identity resolution, data graphs, activations, connections, profile data, ML models, clean rooms, and more — through natural language.

Tools

Data Streams & Segments

| Tool | What it does | |------|-------------| | get_data_streams | List all data streams with status and connector type | | get_segments | List all segments in the org | | publish_segment | Publish / refresh a segment to an activation target |

Data Model Objects (DMOs)

| Tool | What it does | |------|-------------| | get_dmo_schema | Inspect fields on one or more DMOs (filterable by name) | | get_dmo_mapping | Show how a data stream's fields map to a DMO | | propose_dmo_field_mapping | AI-suggested field mappings between a stream and a DMO | | apply_dmo_field_mapping | Write a proposed field mapping back to Data Cloud | | remove_dmo_field_mapping | Remove a field mapping from a DMO |

Data Lake Objects

| Tool | What it does | |------|-------------| | get_data_lake_objects | List all DLOs or fetch a specific one by name/ID (includes fields and schema) |

Data Transforms

| Tool | What it does | |------|-------------| | get_data_transforms | List transforms, get a single one, or fetch its run history / schedule | | upsert_data_transform | Create or update a data transform (SQL or STL definition) |

Calculated Insights

| Tool | What it does | |------|-------------| | get_calculated_insights | List all calculated insights and their SQL | | propose_ci_sql | Generate SQL for a new calculated insight from a description | | create_calculated_insight | Create a new calculated insight in the org | | run_calculated_insight | Trigger a calculated insight to refresh |

Identity Resolution

| Tool | What it does | |------|-------------| | get_identity_resolutions | List IDR rulesets or fetch a specific one (match rules, reconciliation rules) |

Data Graphs

| Tool | What it does | |------|-------------| | get_data_graphs | List data graph metadata, fetch a specific graph, or look up data by entity/ID |

Data Actions & Targets

| Tool | What it does | |------|-------------| | get_data_actions | List Data Actions or Data Action Targets; fetch a specific target by API name |

Activations

| Tool | What it does | |------|-------------| | get_activations | List activations, activation targets, or external platforms; get activation data |

Connections & Connectors

| Tool | What it does | |------|-------------| | get_connections | List connections or connectors; get endpoints, schema, or sitemap for a connection |

Data Spaces

| Tool | What it does | |------|-------------| | get_data_spaces | List data spaces, get a specific one, or list its members |

Profile API

| Tool | What it does | |------|-------------| | get_profile | Fetch profile metadata or query profile records for a data model |

Insight Data

| Tool | What it does | |------|-------------| | get_insight_data | Fetch insight metadata or query calculated insight data with filters/dimensions |

Query

| Tool | What it does | |------|-------------| | query_data | Submit a SQL query, check job status, or fetch result rows |

Search Index

| Tool | What it does | |------|-------------| | get_search_index | List search indexes, get global config, or fetch a specific index |

Machine Learning

| Tool | What it does | |------|-------------| | get_ml_models | List configured models, model artifacts, or model setup versions |

Document Processing

| Tool | What it does | |------|-------------| | get_document_processing | List document processing configurations or fetch global config |

Data Clean Room

| Tool | What it does | |------|-------------| | get_data_clean_room | List collaborations, providers, specifications, or templates; get collaboration jobs |

Data Kits

| Tool | What it does | |------|-------------| | get_data_kits | Get component dependencies or deployment status for a data kit component |

Metadata & Limits

| Tool | What it does | |------|-------------| | get_metadata | Fetch general SSOT metadata (filterable by entity type, name, category, data space) | | get_limits | Get org API limits or current user info |

Prerequisites

A Salesforce Connected App with OAuth scope api and the flow you intend to use enabled (see Auth Methods below).

Auth Methods

client_credentials (default)

Headless machine-to-machine auth. Enable Client Credentials Flow on the Connected App.

Required env vars: SF_CLIENT_ID, SF_CLIENT_SECRET, SF_USERNAME, SF_PASSWORD, SF_LOGIN_URL

jwt

JWT Bearer flow. The user must be pre-authorised on the Connected App. Generate a key pair and upload the certificate to the Connected App.

Required env vars: SF_CLIENT_ID, SF_USERNAME, SF_LOGIN_URL, and either:

• SF_PRIVATE_KEY — PEM content as a string (recommended for hosted/cloud environments)
• SF_PRIVATE_KEY_PATH — absolute path to the .key file (local use only)

Configuration

| Env var | Required | Description | |---------|----------|-------------| | SF_CLIENT_ID | Yes | Consumer key from the Connected App | | SF_LOGIN_URL | Yes | Your org's My Domain URL (e.g. https://yourorg.my.salesforce.com) | | SF_AUTH_METHOD | No | client_credentials (default) or jwt | | SF_CLIENT_SECRET | For client_credentials | Consumer secret from the Connected App | | SF_USERNAME | For client_credentials + jwt | Salesforce username | | SF_PASSWORD | For client_credentials | Salesforce password | | SF_PRIVATE_KEY | For jwt | PEM private key content | | SF_PRIVATE_KEY_PATH | For jwt (local) | Absolute path to .key file | | TOKEN_EXPIRY_BUFFER | No | Seconds before expiry to refresh token (default 300) |

Setup (Claude Desktop — npx)

No install needed. Add to claude_desktop_config.json:

{
  "mcpServers": {
    "salesforce-datacloud-mcp": {
      "command": "npx",
      "args": ["-y", "salesforce-datacloud-mcp"],
      "env": {
        "SF_AUTH_METHOD": "client_credentials",
        "SF_CLIENT_ID": "3MVG9...",
        "SF_CLIENT_SECRET": "71F892...",
        "SF_USERNAME": "[email protected]",
        "SF_PASSWORD": "yourpassword",
        "SF_LOGIN_URL": "https://yourorg.my.salesforce.com"
      }
    }
  }
}

Setup (Cloudflare Workers — remote HTTP)

Deploy your own instance to Cloudflare Workers and connect via HTTP. Credentials are passed as request headers so they stay on your machine.

{
  "mcpServers": {
    "salesforce-datacloud-mcp": {
      "command": "npx",
      "args": [
        "-y", "mcp-remote",
        "https://your-worker.workers.dev",
        "--header", "X-SF-Client-Id:YOUR_CLIENT_ID",
        "--header", "X-SF-Client-Secret:YOUR_CLIENT_SECRET",
        "--header", "X-SF-Username:[email protected]",
        "--header", "X-SF-Password:yourpassword",
        "--header", "X-SF-Login-Url:https://yourorg.my.salesforce.com"
      ]
    }
  }
}

Example prompts

List all data streams in Data Cloud.

Show me all DMOs related to "order".

What fields are on the UnifiedIndividual__dlm DMO?

I have a data stream with fields: email, first_name, last_name, phone.
Suggest how these map to fields on the UnifiedIndividual__dlm DMO.

Create a calculated insight that counts active members per segment.

Troubleshooting

| Symptom | Likely cause | |---------|-------------| | invalid_client_credentials | Wrong SF_CLIENT_ID or SF_CLIENT_SECRET | | invalid_grant | Wrong SF_USERNAME / SF_PASSWORD, or wrong SF_LOGIN_URL | | Data Cloud token exchange failed | Connected App not enabled for Data Cloud; check OAuth scopes | | HTTP 404 on data streams | API version mismatch — verify v63.0 is available in your org | | Tool not visible in Claude | Config JSON syntax error, or Claude Desktop not restarted |