- Home
- MCP servers
- Salesforce Documentation
Salesforce Documentation
- typescript
7
GitHub Stars
typescript
Language
4 months ago
First Indexed
2 months ago
Catalog Refreshed
Documentation & install
Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.
Installation
Add the following to your MCP client configuration file.
Configuration
View docs{
"mcpServers": {
"salesforcediariesbysanket-salesforce-docs-mcp": {
"command": "node",
"args": [
"C:\\path\\to\\salesforce-docs-mcp\\dist\\index.js"
]
}
}
}You have a local-first MCP Server that lets you search Salesforce Developer Documentation offline. It indexes 360 Salesforce PDFs and provides fast, topic-aware search from MCP clients like VS Code or Claude Desktop, while keeping all data on your machine.
How to use
Connect your MCP client to the Salesforce Docs MCP Server and start querying. The server is designed to run locally and respond quickly to intent-based searches across Apex, REST API, LWC, and more.
Typical usage workflow shows these steps: install and build the server, index the PDFs, start the local MCP server, then configure your MCP client to connect via stdio. Once connected, you can perform: search across Salesforce docs, fetch API references, retrieve release notes, and get code examples directly from the indexed documents.
Configure your MCP client to point at the local stdio server. A common setup uses the following command shape to run the server, which you will reference in your client configuration later.
How to install
Prerequisites you need before installation: Node.js 18 or higher, npm or yarn, and an MCP-compatible client (such as a MCP-enabled VS Code or Claude Desktop).
Step-by-step commands to set up the server locally:
# Clone the repository
git clone https://github.com/SalesforceDiariesBySanket/salesforce-docs-mcp.git
cd salesforce-docs-mcp
# Install dependencies
npm install
# Build the TypeScript sources
npm run build
# Build the search index (parses all PDFs - 5-10 minutes)
npm run build-index
Additional configuration and usage notes
The server is designed to run as a local stdio process. You connect to it from your MCP client by specifying the command to start the server and the argument list that launches the compiled index. The following configuration examples demonstrate how to wire the local server into two common clients.
{
"servers": {
"salesforce_docs": {
"type": "stdio",
"command": "node",
"args": ["C:\\path\\to\\salesforce-docs-mcp\\dist\\index.js"]
}
}
}
If you are integrating with Claude Desktop, you can register the same local server using its configuration format and point to the same stdio launcher.
{
"mcpServers": {
"salesforce_docs": {
"command": "node",
"args": ["C:\\path\\to\\salesforce-docs-mcp\\dist\\index.js"]
}
}
}
Security and data handling
All data stays on your machine. There are no API keys required and no external HTTP endpoints exposed. The server uses parameterized queries and input validation to protect against invalid input.
To ensure offline operation and local control, index your PDFs locally and run the server without sending data to remote services.
Notes and troubleshooting
If you encounter slower startup or indexing times, ensure the index build completed successfully and that the PDFs are placed under docs/pdfs as described in the project structure. The index build may take several minutes depending on your machine.
For client configuration, verify that your MCP client is set to communicate with the local stdio process using the exact command and arguments shown in the example blocks. Replace the path with your actual installation path.
Available tools
search_salesforce_docs
Search across all Salesforce documentation with natural language queries.
get_api_reference
Get specific Salesforce API reference documentation.
get_release_notes
Get Salesforce release notes for specific releases or features.
get_code_example
Get code examples from Salesforce documentation.
list_doc_categories
List all documentation categories with document counts.
get_document
Get full content of a specific document by ID or filename.