Ganty Now Supports MCP: Drive Your Gantt Charts Directly From Claude

Today, Ganty officially supports the Model Context Protocol (MCP). You can now read and write Gantt chart tasks from Claude Desktop, Claude Code, Cursor, and other MCP-compatible AI assistants — using natural language. Setup takes three minutes, and it's free on every plan. This post covers what MCP is, why we built it, what you can do, and how to get started.

What Is MCP?

Released by Anthropic in 2024, MCP is an open-source protocol that provides a unified language between AI assistants like Claude and external services or data sources. Previously, integrating AI with a SaaS product required building separate plugins or extensions for ChatGPT, Claude, Cursor, and so on. With MCP, you implement one MCP server and every compatible AI client can use it.

Just as USB unified connections to peripheral devices with a single cable standard, MCP aims to unify how AI connects to services. Adoption accelerated through 2025, and today it's supported by Claude Desktop, Claude Code, Cursor, Zed, Windsurf, and many other AI tools.

Why Ganty Built MCP Support

1. Project Management Is a Natural Fit for AI

The actual work of project management — identifying tasks, organizing dependencies, gathering progress, writing reports — is endlessly about consolidating, organizing, and updating information. That's exactly what AI does well. Yet until now, the workflow was "work in Ganty → copy into ChatGPT → paste results back into Ganty." With MCP, AI accesses Ganty directly.

2. A Clear Industry Differentiator

Major Gantt chart SaaS products (Asana, Smartsheet, Microsoft Project, ClickUp) don't yet offer official MCP servers. Ganty is one of the first dedicated Gantt SaaS products to ship native MCP support, which we believe will be a meaningful advantage as AI-native project workflows become standard.

3. The Experience Users Actually Want

Users don't want to "build a Gantt chart." They want to "move their project forward." Asking Claude to "organize the tasks we need before next week's launch" delivers value that traditional UIs simply cannot.

What You Can Do Today

Ganty MCP exposes 13 tools.

Read tools

• list_workspaces - your workspaces
• list_projects - projects in a workspace
• list_tasks - tasks in a project (with search/status/assignee filters)
• get_task - detailed task info including assignees and dependencies
• list_milestones - project milestones

Write tools

• create_project - create a new project
• delete_project - delete a project
• create_task - create a new task
• update_task - update any task fields
• set_task_progress - update progress percentage (status auto-derived)
• add_dependency - add a finish-to-start dependency
• delete_task - delete a task
• create_milestone - add a milestone

Three Real Scenarios

Scenario 1: Generate an Entire Project

You: "Create a 3-month e-commerce redesign project. Tasks should flow: requirements → design → frontend → backend → QA → launch, with proper dependencies."

Claude (10 seconds later): "Done. Project created with 6 tasks, dependencies wired up, launch milestone set. Want to review?"

Building a Gantt chart from scratch now takes under a minute.

Scenario 2: Bulk Progress Updates

You: "Update all design-related tasks to 70%."

Claude: "Found 3 matching tasks: UI Design, Logo Design, Banner Design. All updated to 70%."

What used to require navigating between 3 task panels now happens in one message.

Scenario 3: Project Health Analysis

You: "Which tasks must finish this week to keep the critical path?"

Claude: "The 'API Integration' task in Project A. Currently 40% complete with 3 days left. Two downstream tasks depend on it, so a slip here pushes the launch back by a week."

AI looks at the whole project graph and surfaces risks humans miss. This is the real promise of AI project management.

Three-Minute Setup

Step 1: Generate an API Token

Sign in to Ganty → top-right avatar → "API Tokens" → name it and create. Copy the gnty_xxx... token immediately (it cannot be shown again).

Step 2: Configure Your MCP Client

For Claude Desktop, add to claude_desktop_config.json:

{
  "mcpServers": {
    "ganty": {
      "type": "http",
      "url": "https://ganty.app/api/mcp",
      "headers": {
        "Authorization": "Bearer gnty_xxx..."
      }
    }
  }
}

For Claude Code, run:

claude mcp add --transport http ganty https://ganty.app/api/mcp \
  --header "Authorization: Bearer gnty_xxx..."

Step 3: Talk to Claude

Try "List my Ganty workspaces" or "Show the progress of last week's project." Full setup details on the MCP integration guide.

Why Free on Every Plan

We could have made MCP a Pro/Business-only feature. Most competitors gate API access behind paid tiers. We chose free everywhere because:

• The AI-native PM experience should reach everyone. It's too new to gate by price; we want feedback from the entire user base.
• It strengthens the free plan's position. "Free Gantt chart with AI integration" is a compelling acquisition message.
• Usage drives natural upgrade demand. Heavy MCP users hit task and member limits sooner, creating organic upgrade signals.

Security

API tokens are stored as SHA-256 hashes — even Ganty admins cannot recover the raw value. Tokens are shown exactly once at creation time, and can be revoked instantly from settings if compromised. On the Business plan, all write operations are recorded in the audit log for full traceability.

Roadmap

Starting with 13 tools, we'll add the following based on user feedback:

• Auto-assign by email lookup
• Share link creation and revocation
• Audit log queries
• Report generation (auto-summary of project status)
• OAuth support (SSO flow, no manual token management)

Related Reading

• How AI Is Transforming Project Management
• Multi-Period Tasks
• Why Excel Gantt Charts Break Down

Try It Today

MCP works on the free plan starting today. Claude Desktop users can be up and running in three minutes. Experience the shift from "drawing Gantt charts" to "running projects with AI as your co-pilot." Read the full integration guide.