MCP Server
|
The ITM Platform MCP server connects AI assistants and agents to your portfolio through the Model Context Protocol. Your AI client can search projects, inspect budgets, summarize portfolio health, create tasks, log risks and issues, and update project details, always using your own ITM Platform permissions. |
It works with Claude, VS Code, Cursor, OpenAI Codex, Windsurf, JetBrains AI Assistant, and any other client that supports MCP.
The full technical documentation, including client-by-client setup and the complete tool reference, is available at the ITM Platform MCP Documentation on the ITM Platform Developers site.
What is MCP?
The Model Context Protocol (MCP) is an open standard that lets AI assistants connect to external systems in a consistent way. Instead of copying and pasting data into a chat, you connect your AI client to the ITM Platform MCP server once, and the assistant can then read and act on your portfolio data on your behalf.
The server authenticates as you, calls the ITM Platform APIs, and returns only the data your ITM Platform account is allowed to access.
What can an agent do?
From simple lookups to fully automated workflows, MCP unlocks progressively more powerful use cases:
- Quick lookups -- Ask a question and get an answer, such as "What risks are open across my portfolio?"
- Multi-step analysis -- The agent chains several tools and synthesizes results, for example reviewing every project ending this quarter and flagging budget overruns, high-impact risks, or low task completion.
- Automated bulk actions -- The agent reads, decides, and writes across projects, such as moving stalled projects to Execution and creating a kick-off task for each project manager.
- Scheduled intelligence -- An agent runs on a schedule with no human prompt, for example pulling overdue tasks every Monday and posting a summary grouped by project manager.
- Cross-system orchestration -- Combine ITM Platform's MCP with other MCP servers (GitHub, Slack, calendar, email) so an action in one system triggers updates in ITM Platform.
What can it access?
The server exposes a set of tools for reading and updating your data, plus resources and guided prompts for common analysis tasks.
Read -- Search and retrieve projects, services, tasks, budgets, purchases, revenues, risks, issues, users, and reference data, and run aggregated portfolio and DataMart queries for advanced analysis.
Write -- Create tasks, risks, and issues, and update tasks and project details such as status, dates, assignees, and priority.
Guided prompts -- Ready-made workflows for project status summaries, portfolio overviews, team workload reviews, and risk analysis.
For the complete list of tools, resources, and prompts, see the MCP Documentation.
How to connect
There are two ways to connect your AI client. Both are covered step by step in the developer documentation.
| Method | Connection value | Best for |
|---|---|---|
| Hosted (OAuth) | https://api.itmplatform.com/v2/_/mcp/ |
Most users and managed AI clients. Nothing to install, just sign in with your ITM Platform account. |
| Local (API key) | npx @itm-platform/mcp-server |
Local execution, firewalled networks, or self-hosted ITM Platform instances. |
With the hosted option, your AI client never sees your ITM Platform password or API key: you sign in through ITM Platform and approve the requested access. This is the recommended setup for most users.
With the local option, the server runs on your machine and authenticates with a personal API key. To create one, log in to ITM Platform, open My Profile, and generate a key from the API Key section.
After adding either connection, open your client's MCP command or server list. In clients that support slash commands, type /mcp, select itm-platform, and authenticate when prompted.
Who can access it?
The MCP server uses the same identity and permission model as ITM Platform, so an agent can only see and change what you could see and change yourself.
| License | MCP access |
|---|---|
| Company Administrator | Full read and write access |
| Full User | Full read and write access |
| Project Manager | Not yet available |
| Team Member | Blocked |
When connecting with OAuth, access is further controlled by scopes: mcp:read allows read-only tools, while mcp:write adds create and update tools.
Project data is returned to the AI client you choose, so the AI provider's data-handling policy applies to any data it processes. All AI-generated answers should be reviewed before making decisions.
Getting started
The full technical documentation, including client-specific setup, the complete tool and prompt reference, authentication details, and troubleshooting, is available at the ITM Platform Developers site:
ITM Platform MCP Documentation
Related
- DataMart -- read-optimized GraphQL data layer used by some MCP analysis tools.
- API -- the REST API the MCP server calls on your behalf.
- Microsoft Teams Bot -- ask PMPilot about your ITM Platform data directly in Teams.