The Next Evolution of Salesforce Testing
Salesforce quality engineering is inherently complex. Page Objects, layered test hierarchies, coverage gaps, and CI/CD orchestration all demand ongoing attention. Managing this manually is time-intensive and introduces risk.
Provar MCP changes that dynamic.
Now available in beta at no additional cost, Provar MCP introduces an AI-assisted quality layer directly within the Provar DX CLI. Built on the Model Context Protocol (MCP), it enables AI assistants like Claude and Cursor to take real action inside your test projects. The outcome is more efficient test creation, clearer quality insights, and streamlined execution.
What Is the Model Context Protocol (MCP)?
The Model Context Protocol is an open standard that allows AI assistants to interact with tools on your behalf.
Instead of generating static suggestions, MCP-enabled AI can:
- Read project files
- Execute commands
- Generate code
- Return real-time results
For Salesforce testing teams, this shifts AI from advisory to operational. Provar MCP exposes core project functions to AI clients, translating natural language into concrete actions within your testing environment.
What Can Provar MCP Do?
Once configured, Provar MCP allows AI to operate across your testing lifecycle without leaving a single interface:
Project visibility and coverage analysis
Quickly assess test case counts, suite structures, Page Objects, and uncovered gaps. Tasks that once required manual audits are surfaced immediately.
Test case and Page Object generation
Generate Java Page Objects with correct annotations and locator strategies, or create or edit Provar test cases with proper structure and identifiers. Test steps can be built dynamically, from navigation to field input. Provar MCP also supports the generation, discovery, and maintenance of NitroX components used by Provar tests, ensuring compatibility with the latest Provar artifacts from the start.
Quality validation at every level
Validate text cases, suites, plans, and full projects against our quality engine, boasting over 170 quality rules built in.
Run configuration management
Create, update, and validate your provardx-properties.json file to ensure consistent execution environments, as well as Provar ANT files used for test execution.
End-to-end execution
Trigger local runs through Provar Automation or managed runs via Provar Quality Hub, including compile, execute, and reporting workflows.
Root Cause Analysis and defect creation
Classify failures by type, identify impacted Page Objects, and convert failed tests into Provar Quality Hub defects directly from the same session.
Security By Design
Provar MCP is built with a strict local-first security model.
- The MCP server runs entirely on your machine
- No network TCP ports are opened
- File access is limited to approved project paths
- Path traversal is blocked
- All files operations are strictly scoped to the project directory you specify via the –allowed-paths flag
Importantly, no test code, project files, or credentials are transmitted externally. Salesforce credentials remain within the Salesforce CLI’s credential store and are never accessed by the MCP server.
Every interaction is logged with a unique request ID, providing a complete audit trail.
Getting Started with Provar MCP
For teams already using Provar Automation and the Salesforce CLI, setup is straightforward.
Requirements include:
- Provar Automation (v2.18.2 or 3.0.6+) with an activated license
- Salesforce CLI (v2.x or later)
- ProvarDX CLI plugin (v1.5.0-beta or later)
- Node.js 18+
- An MCP-compatible AI client (e.g., Claude Desktop, Claude Code, Cursor)
Installation involves just three steps:
- Installing the Salesforce CLI
- Adding the ProvarDX CLI plugin
- Updating your AI client configuration with a short JSON configuration block (For Claude Desktop, this means editing claude_desktop_config.json; for Cursor, it’s done through the MCP settings panel.).
Once configured, Provar tools appear directly within your AI interface. A simple ping command confirms connectivity.
A Practical Workflow Example
A typical AI-assisted workflow with Provar MCP might look like:
- Inspect the project: “Tell me how many test cases exist and which ones aren’t covered by any test plan.”
- Generate missing coverage: “Generate a test case called ‘Verify Account Creation’ and write it to the smoke test folder.”
- Validate quality: “Validate the full project and give me a quality report with per-plan scores.”
- Run the tests: “Load the properties file, compile the project, run the tests, and tell me the results.”
- Triage failures: “Classify the failures in the Results folder and tell me which look like new regressions.”
- Log defects: “Create a Provar Quality Hub defect for the failed LoginTest.”
What previously required multiple tools and manual coordination now happens within a single, continuous interaction.
Who Should Use Provar MCP
Provar MCP is especially valuable for:
- Salesforce QA engineers scaling automation efforts
- Salesforce developers wanting test automation as part of their AI workflows
- DevOps teams integrating testing into CI/CD pipelines
- Organizations managing large or legacy Provar projects
- Teams exploring structured, governed AI adoption in testing
Because Provar MCP is open source and included for existing Provar Automation users, adoption does not introduce additional investment barriers.
AI and the Future of Salesforce QA
Salesforce automated testing is moving beyond validation toward intelligent, action-oriented systems.
Provar MCP reflects that shift. By combining Salesforce-specific quality rules with Provar MCP’s flexible execution model, teams can embed AI into their workflows in a way that is both controlled and productive.
As the product moves toward general availability, early adopters will be better positioned to operationalize AI within their quality practices.
Bring AI Into Your Testing Workflow
Provar MCP offers a practical path to AI-assisted testing without sacrificing control or visibility.
If you are evaluating how AI fits into your Salesforce testing strategy, Provar MCP is a strong place to start.
Ready to bring AI into your Salesforce test automation workflow? Reach out to Provar to explore.
