Skip to main content

MCP Integration

Pidima exposes the engineering lifecycle to AI tools and agents through the Model Context Protocol (MCP). Connect your IDE assistant or internal agents to Pidima and they gain governed access to your project's engineering context:

  • Requirements — content, hierarchy, status and revision history
  • Architecture — elements linked to each requirement
  • Verification — linked test cases, latest results and coverage
  • Change — what changed recently and what it impacts

Access through MCP respects the same project permissions, roles and audit trail as the Pidima application — an agent can only see and do what its user can.

Product Review Required

[PRODUCT REVIEW] This page requires additional technical details:

  • Supported clients
  • Endpoint/setup instructions
  • Authentication method
  • Worked example (e.g., asking the IDE assistant "why does this requirement exist?")

Why Use MCP?​

MCP (Model Context Protocol) is an open standard that enables AI assistants to securely connect to external systems. With Pidima's MCP integration, your AI tools can:

  • Query requirements and understand project context
  • Check verification status and test coverage
  • Understand architecture relationships
  • Track changes and their impact across the lifecycle

Getting Started​

For detailed setup instructions including API key creation and client configuration, see the MCP Server Integration page.

Use Cases​

IDE-Assisted Development​

Connect your IDE (Cursor, VS Code with AI extensions) to Pidima to:

  • Ask "why does this requirement exist?" and get full context including linked documents
  • Query "what tests cover this functionality?" while reviewing code
  • Understand "what changed in the last sprint that affects this component?"

Agent-Based Workflows​

Build internal automation agents that:

  • Generate reports on verification coverage
  • Alert on requirements that need attention
  • Cross-reference code changes with requirements

AI-Powered Analysis​

Let AI assistants help with:

  • Impact analysis across requirements and tests
  • Gap identification in test coverage
  • Compliance checking against uploaded standards

Security and Permissions​

All MCP access is governed by the same security model as the Pidima web application:

  • Authentication: Uses API keys tied to user accounts
  • Authorization: Agents can only access projects the user has permission to view
  • Audit Trail: All queries and actions are logged
  • Data Boundaries: No data leaves your Pidima instance through MCP