What it does

Exposes Salesforce orgs to an agent: describe objects, query records, and inspect metadata. The hard part is never the connection, it is the permission model — an agent sees exactly what its integration user sees, which is often everything. Point it at a sandbox and a least-privilege user before anything else.

Install

Terminal

npx -y @salesforce/mcp

Client configuration

Paste this into your MCP client config (Claude Desktop, Cursor, or any client that reads mcpServers), then restart the client.

mcp.json

{
  "mcpServers": {
    "salesforce": {
      "command": "npx",
      "args": [
        "-y",
        "@salesforce/mcp"
      ]
    }
  }
}

Use it for

Querying org metadata and records during a CRM integration

Skip it if

Production orgs without a sandbox first; Salesforce permissions are the real limit

When to use it

Reach for this when the task needs crm inside the agent loop. Skip it if a plain shell command or a one-off script is enough — every server you add costs context and adds a failure mode.

Other TypeScript servers

Filesystem — Gives an agent scoped read, write, and search access to directories you explicitly allow
Memory — A persistent knowledge graph stored in a local JSON file, exposed through entities, relations, and observations
Sequential Thinking — Gives the model a scratchpad for structured, revisable reasoning: numbered thoughts that can branch, revise, and conclude

Salesforce MCP