Forecast AI: Getting Started

This guide walks you through both ways to access the Forecast MCP Server, along with the safeguards that keep you in control, so you can get started with confidence. 

Forecast AI lets you use your preferred AI assistant to get answers and take action on your Forecast data. You can use it two ways: 

• Through Ask AI, built into Forecast with nothing to set up

• By connecting your preferred AI assistant (such as Claude or ChatGPT) via the Forecast MCP server.

Either way, you can ask about your projects, tasks, and schedules and take action by reassigning work, rescheduling tasks, and logging time, all through a natural conversation.

Jump to...

• How to access 
  • In the Forecast application through “Ask AI.”
  • Connect your own LLM (Large Language Model)
    • Claude 
    • ChatGPT
    • Perplexity
    • Grok
    • Mistral Le Chat
    • Microsoft Copilot Studio
• Built-in safeguards 

How to access

In the Forecast application through “Ask AI.”

Click “Ask AI” in the primary navigation at the top-right corner of your screen. The assistant is built in, so there is nothing to set up.

Connect your own LLM (Large Language Model)

Use your preferred AI assistant by connecting the Forecast MCP server.

Key things to know:

• Because MCP is an open standard, it isn't tied to a single AI tool. If your organization already uses a compatible AI assistant, you can add Forecast data to your existing setup instead of adopting a new one. 
• You can combine Forecast data with other sources connected to your LLM, such as email, CRM, or accounting systems, in a single conversation. 
• When you set up a custom connector, you can name it whatever you like. Throughout this section of the help guides, Forecast MCP is used as the example name, but your connector may have a different name.
• Every request uses your signed-in Forecast permissions, so the same access rules and controls from the application always apply, and every change is shown to you before it's applied.

Claude

To enable this for your organization, a Claude admin must complete the initial setup in Claude by going to Organization Settings > Connectors: 

1. Go to claude.ai
2. Select Customize.
3. Select Connectors.
4. Click the + next to Connectors, then select Add custom connector.
5. In the "Add custom connector" modal:
   • Name: Forecast MCP 
   • Remote MCP Server URL: https://mcp.forecast.it/mcp
6. Click Add.
7. Click Connect, then follow the prompts to authenticate and authorize your Forecast account using email login or SSO.
8. Return to Claude, and you're ready to go.
9. Once enabled for your organization, anyone with access to your organization’s Claude account can connect Claude to the Forecast MCP server.

After this has been enabled for your organization, team members can connect their Forecast account by following these steps:

1. Go to their Claude account.
2. Go to Customize > Connectors and find Forecast MCP in the list.
3. Click Connect, then follow the prompts to authenticate and authorize the Forecast connection using email login or SSO.

ChatGPT

Custom connectors are accessed through ChatGPT’s Developer Mode. This is a per-user setup, meaning each user completes the steps below individually. 

A paid ChatGPT plan (Plus or above) is required, and for Team/Enterprise plans, a ChatGPT admin must first enable Developer Mode for the organization.

1. Navigate to ChatGPT .
2. Go to Settings.
3. Select Apps & Connectors.
4. Open Advanced settings at the bottom and enable Developer mode (if allowed by your organization).
5. Return to Apps & Connectors and click Create.
6. In the create window:
   • Enter Forecast MCP as the Name.
   • Add a short Description.
   • Enter https://mcp.forecast.it/mcp in the MCP Server URL field.
   • If prompted, select I trust this application.
7. Click Create.
8. Follow the prompts to authenticate and authorize your Forecast account using email login or SSO.

To use Forecast MCP in a chat:

1. Start a new conversation.
2. Click the + near the message box.
3. Select More and choose Forecast MCP.

• Forecast MCP is enabled per conversation, so you may need to select it again when starting a new chat.
• ChatGPT will ask you to manually confirm any action that creates, updates, or changes data in Forecast.

Perplexity

Custom remote connectors require a paid plan - available on Pro, Max, and Enterprise; launched March 13th, 2026. Individual Pro/Max users can add the connector directly. 

For Enterprise/organization accounts, an admin must first enable member-added connectors (this is off by default), the same way Claude requires org-level setup first.

Truthifi

Org enablement (Enterprise admins only):

1. Go to perplexity.ai and sign in as an admin.
2. Go to Enterprise settings → Permissions → Connectors permissions.
3. Turn on Allow members to add custom connectors (and/or add it organization-wide yourself).

To add the connector (per user):

1. Go to perplexity.ai.
2. Open Account settings → Connectors.
3. Click + Custom connector in the top-right corner.
4. In the pop-up, select Remote.
5. Fill in the fields:
   • Name: Forecast MCP
   • MCP Server URL: https://mcp.forecast.it/mcp
   • Authentication: OAuth
   • Transport: Streamable HTTP
6. Check the acknowledgement box to confirm you understand the risks of custom connectors.
7. Click Add.
8. Click the connector card to start the authentication flow, then follow the prompts to authenticate and authorize your Forecast account using email login or SSO.

To use Forecast MCP in a thread:

1. Click the + icon in the query box.
2. In the Connectors and sources menu, confirm Forecast MCP is selected.

Grok

Custom MCP support ("Bring Your Own MCP") requires a paid Grok account and is a per-user setup. It launched May 6th, 2026 on Grok Web, iOS, and Android. 

For Grok Business/Enterprise teams, a team admin provisions the connector from the cloud console first, after which members connect their own accounts. 

GitHub

1. Go to grok.com and sign in.
2. Open the left-hand Menu and click Connectors (or go directly to grok.com/connectors).
3. Click New Connector, then select Custom. (Depending on the build, this button may read "Add custom connector.")
4. Enter the details:
   • Name: Forecast MCP
   • MCP Server URL: https://mcp.forecast.it/mcp
   • Add a short Description if prompted.
5. Confirm the risk acknowledgement.
6. Follow the prompts to authenticate and authorize your Forecast account using email login or SSO. Grok will discover the tools your MCP server exposes and make them available in conversations. xAI Docs

To use Forecast MCP in a chat:

1. Start a new conversation.
2. Click the tools/connectors icon in the composer and toggle Forecast MCP on.

• The connector is account-scoped, so once added it appears across every Grok session you sign into. 
• Team admins provisioning for an organization do so from the console Connectors page via + Add Connector; this only creates the connector definition — each member still runs their own OAuth flow.

Mistral Le Chat

Connectors are available to all Le Chat users, including the Free plan. 

Creating a custom connector requires Admin privileges, but Free, Pro, and Student account owners are Admin by default. For Team/Enterprise accounts, an admin may need to enable AI Connectors first. 

Mistral AImistral

1. Go to chat.mistral.ai and sign in.
2. Click the toggle panel button to reveal the side panel.
3. Expand the Intelligence menu, then click Connectors.
4. Click + Add Connector, then select the Custom MCP Connector tab.
5. Enter the details:
   • Name: Forecast MCP
   • Server URL: https://mcp.forecast.it/mcp
   • Description (optional)
   • Authentication: OAuth 2.1
6. Click Connect. Le Chat validates the connection and guides you through the OAuth consent flow to authenticate and authorize your Forecast account using email login or SSO.
7. Once successful, Forecast MCP appears in your list of connectors, ready to enable.

To use Forecast MCP in a chat: click the enable-tools icon and ensure Forecast MCP is selected.

Built-in safeguards: 

• Permissions & Access
  • Every request uses your signed-in Forecast permissions, so the same access rules and controls from the application always apply, and every change is shown to you before it's applied.
• Review Changes Before Applying
  • The assistant shows you exactly what it’s about to change and waits for your approval before making any updates.
• Choose Which Changes to Apply
  • When applying multiple changes at once, you can deselect anything you want to skip and confirm only the changes you want to make.
• Preventing Scheduling Conflicts
  • You’ll be alerted if a change extends work beyond a project’s end date or if a rescheduled allocation conflicts with another assignment.

Keep exploring

• Forecast MCP Server — An introduction to Forecast MCP and what it can do for you. 
• Forecast AI: Capabilities and Prompt Examples — Explore the full list of questions you can ask and actions you can take, with example prompts.
• How Forecast's AI Features Use Your Data — This article explains how Forecast's AI features use your data.