salesforce-mcp
Provides read-only access to Salesforce orgs via MCP, enabling SOQL queries, sObject descriptions, and object listing using OAuth 2.0 Client Credentials.
README
Salesforce MCP Server (Read-Only)
An MCP server that provides read-only access to your Salesforce org: run SOQL queries, describe sObjects, and list all objects. Authentication uses the OAuth 2.0 Client Credentials flow. No writes or DML are allowed -- only SELECT queries and describe/list APIs.
Supports two transport modes:
- stdio -- for local use with Claude Desktop
- Streamable HTTP -- for remote deployment (Docker, Azure Container Apps, managed Claude agents)
Requirements
- Python 3.10+
- A Salesforce Connected App configured for the Client Credentials flow
Connected App Setup (Salesforce)
- In Salesforce: Setup > App Manager > New Connected App.
- Enable OAuth Settings.
- Under Selected OAuth Scopes, add Access and manage your data (api).
- Enable Client Credentials Flow.
- Save and note your Consumer Key and Consumer Secret.
- Under Manage > Edit Policies, set the Client Credentials run-as user.
Environment Variables
| Variable | Required | Description |
|---|---|---|
SALESFORCE_CLIENT_ID |
Yes | Connected App Consumer Key |
SALESFORCE_CLIENT_SECRET |
Yes | Connected App Consumer Secret |
SALESFORCE_LOGIN_HOST |
No | login.salesforce.com (default) or your My Domain host |
MCP_HOST |
No | Server bind address (default: 0.0.0.0) |
MCP_PORT |
No | Server port (default: 8765) |
MCP_API_KEY |
No | API key for Bearer token auth on the /mcp endpoint. When unset, auth is disabled. |
MCP_ALLOWED_HOSTS |
No | Comma-separated allowed Host headers, or * to disable DNS rebinding protection (recommended for cloud deployments with API key auth). |
MCP_TRANSPORT |
No | Transport mode when no CLI arg is given (stdio or streamable-http, default: stdio). |
Running Locally (stdio)
pip install .
python -m salesforce_mcp
This starts the server in stdio mode, suitable for Claude Desktop.
Claude Desktop Configuration
{
"mcpServers": {
"salesforce": {
"command": "python",
"args": ["-m", "salesforce_mcp"],
"env": {
"SALESFORCE_CLIENT_ID": "<your_consumer_key>",
"SALESFORCE_CLIENT_SECRET": "<your_consumer_secret>",
"SALESFORCE_LOGIN_HOST": "<your_login_host>"
}
}
}
}
Running Locally (Streamable HTTP)
python -m salesforce_mcp streamable-http
The server starts on http://0.0.0.0:8765/mcp with a health check at /health.
Running with Docker
docker compose up --build
Pass credentials via a .env file in the project root. The Dockerfile runs the server in Streamable HTTP mode by default.
Deploying to Azure Container Apps
-
Build and push the Docker image to your Azure Container Registry:
az acr login --name <your_acr> docker build --platform linux/amd64 -t <your_acr>.azurecr.io/salesforce-mcp:latest . docker push <your_acr>.azurecr.io/salesforce-mcp:latest -
Create the container app:
az containerapp create \ --name salesforce-mcp \ --resource-group <your_rg> \ --environment <your_cae> \ --image <your_acr>.azurecr.io/salesforce-mcp:latest \ --registry-server <your_acr>.azurecr.io \ --target-port 8765 \ --ingress external \ --min-replicas 0 \ --max-replicas 3 \ --cpu 0.25 \ --memory 0.5Gi \ --env-vars \ SALESFORCE_CLIENT_ID=secretref:salesforce-client-id \ SALESFORCE_CLIENT_SECRET=secretref:salesforce-client-secret \ SALESFORCE_LOGIN_HOST=<your_login_host> \ MCP_HOST=0.0.0.0 \ MCP_PORT=8765 \ MCP_API_KEY=secretref:mcp-api-key \ MCP_ALLOWED_HOSTS="*" \ --secrets \ salesforce-client-id="<your_client_id>" \ salesforce-client-secret="<your_client_secret>" \ mcp-api-key="<your_api_key>" -
Connect your MCP client to the deployed endpoint:
{ "mcpServers": { "salesforce": { "type": "streamable-http", "url": "https://<your_fqdn>/mcp", "headers": { "Authorization": "Bearer <your_api_key>" } } } }
API Key Authentication
When MCP_API_KEY is set, all requests to /mcp must include an Authorization: Bearer <key> header. Requests without a valid key receive a 401 Unauthorized response.
The /health endpoint is always unauthenticated so that platform health probes (Azure, Docker, etc.) work without credentials.
When MCP_API_KEY is not set, all requests pass through without auth -- suitable for local development.
Tools (Read-Only)
| Tool | Description |
|---|---|
| run_soql | Execute a SOQL SELECT query. INSERT/UPDATE/DELETE/UPSERT/EXECUTE are rejected. |
| describe_sobject | Describe one sObject: fields, labels, types, relationships. |
| list_objects | List all sObjects in the org: name, label, custom flag. |
License
MIT.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.