powerplatform mcp/cli

What is powerplatform mcp/cli?

A Model Context Protocol (MCP) server and standalone CLI for querying and configuring PowerPlatform / Dataverse environments. It supports multiple environments, entity metadata, records, plugins, flows, solutions, workflows, business rules, security roles, custom APIs, web resources, and write operations for automated environment setup. It is designed for developers and administrators using AI-powered clients like Claude, Cursor, and GitHub Copilot.

How to use powerplatform mcp/cli?

Install via npm (npm install -g powerplatform-mcp or npx powerplatform-mcp for the MCP server; npm install -g powerplatform-cli or npx powerplatform-cli for the CLI) or via Docker. Configure multiple environments using environment variables (e.g., POWERPLATFORM_DEV_URL, POWERPLATFORM_DEV_CLIENT_ID). The MCP server exposes 67 tools that accept an optional environment parameter; the CLI writes results to a file system cache to avoid context window limits.

Key features of powerplatform mcp/cli

• Query and configure Dataverse environments via MCP or CLI.
• Supports 67 tools covering entities, records, plugins, flows, solutions, security roles, and more.
• Multiple environment support with environment-specific credentials.
• Write operations for automated environment setup (e.g., create attributes, plugin steps, security roles).
• CLI persists full results to disk to avoid AI context window truncation.
• Docker images available for both MCP server and CLI.

Use cases of powerplatform mcp/cli

• Conversational exploration of PowerPlatform environments using AI clients.
• Automated environment setup and configuration via scripts or AI assistants.
• Bulk analysis of flows, plugins, solutions, and security roles.
• Troubleshooting using plugin trace logs and flow run history.
• Exporting solutions and managing solution components.

FAQ from powerplatform mcp/cli

What is the difference between the MCP server and the CLI?

The MCP server integrates with AI clients for interactive use, returning results inline. The CLI writes results to a file system cache to avoid context window limits, persisting full results for follow-up analysis.

What are the runtime requirements?

Requires Node.js 22+ (less than 25).

How do I configure multiple environments?

Define environment variables: POWERPLATFORM_ENVIRONMENTS=DEV,UAT,PROD and for each environment set POWERPLATFORM_<ENV>_URL, _CLIENT_ID, _CLIENT_SECRET, and _TENANT_ID.

What transport does the MCP server use?

The README does not specify the transport protocol (e.g., stdio, HTTP). It is designed for AI-powered clients but does not detail transport.

Are there any known limits?

The CLI avoids AI context window limits by writing to disk. The query-records tool has a default maxRecords of 50. Other tools mention maxRecords optional parameters.