HighLevel Project Management now connects with compatible AI assistants through Model Context Protocol (MCP), making it easier to organize and manage project work using natural language. With this integration, you can create and update tasks, subtasks, Spaces, and Lists without switching between your AI client and HighLevel. You can also work with custom statuses, custom field values, and task comments through an authorized connection. This guide covers how to connect a supported AI client using OAuth and start managing Project Management work through MCP.

TABLE OF CONTENTS

• What is Project Management MCP?
• Key Benefits of Project Management MCP
• How To Setup Project Management MCP
• Connect Your AI Client
  • Authorize Your Connection with OAuth
  • Send Your First Project Management Request
• Manage Project Work Using Natural Language
  • Create and Update Tasks, Subtasks, Spaces, and Lists
  • Work with Custom Statuses and Fields
  • Read and Post Task Comments
• Frequently Asked Questions
  • Related Articles

What is Project Management MCP?

Project Management MCP is an integration that connects compatible AI clients to HighLevel Project Management through Model Context Protocol (MCP). MCP allows an AI assistant to discover and use supported HighLevel operations through a standardized connection.

Instead of opening Project Management whenever you need to create tasks, organize Lists, or review comments, you can describe the action in your connected AI client.

The integration supports creating and managing tasks, subtasks, Spaces, and Lists. It also supports working with custom statuses, custom field values, and task comments.

Connections use a hosted MCP endpoint and OAuth authentication. You authorize the connection through your own account rather than configuring shared credentials or building directly against the Projects REST API.

Supported AI Clients

• Claude.ai

• ChatGPT

• Cursor

• Codex

• Claude Code

• Muse

Key Benefits of Project Management MCP

Project Management MCP helps you manage project work from your AI assistant while keeping that work inside HighLevel. Natural-language instructions reduce routine navigation and provide another way to interact with project information.

• Natural-language task management: Create and update tasks, subtasks, Spaces, and Lists by describing the work you want to complete.

• Less context switching: Manage supported project activities without repeatedly opening the Project Management interface.

• Custom statuses and fields: Create tasks with appropriate statuses and custom field values, reducing manual adjustments after creation.

• Task discussions: Read task comments and post new comments through your connected AI client.

• Simplified integration: Use hosted MCP connectivity to work with supported project operations without building directly against the Projects REST API.

• OAuth-based authentication: Connect through your own authorization and grant access to the requested HighLevel resources.

• No local server setup: The hosted endpoint does not require a locally installed MCP server or VPN connection.

How To Setup Project Management MCP

Connecting Project Management to an AI assistant involves configuring a hosted MCP connection and completing OAuth authorization. Once connected, you can use natural-language requests to work with supported project resources.

The initial connection process starts inside your AI client. HighLevel authentication takes place when the client begins the OAuth authorization flow.

Connect Your AI Client

Choose the MCP endpoint associated with your AI client. The following endpoints are documented in HighLevel's MCP resources.

Connect ChatGPT

ChatGPT supports a dedicated OAuth-based MCP connection. You can configure the connection as a custom MCP app when your ChatGPT account and workspace permissions allow it.

1. Open ChatGPT on the web.

2. Open Settings.

3. Select Apps.

4. If your account requires Developer mode, enable it before creating the MCP connection.

5. Select Create to begin creating a custom MCP app.

6. Enter a name for the connection, such as HighLevel Project Management.

7. Add a description identifying the connection as your HighLevel MCP integration.

8. Enter the following MCP server URL:

   https://services.leadconnectorhq.com/mcp/openai/v2/

9. Continue with the MCP app configuration.

10. Select the available option to scan or discover the MCP tools.

11. Begin OAuth authorization when prompted.

12. After completing authorization, return to ChatGPT and finish creating the app.

Connect Claude.ai

Claude.ai supports HighLevel MCP connections through a custom connector. After adding the hosted endpoint, complete OAuth authorization to make the approved HighLevel operations available.

1. Open Claude.ai.

2. Select Settings.

3. Select Connectors.

4. Click Add custom connector.

5. Enter the following server URL:

   https://services.leadconnectorhq.com/mcp/anthropic/v2

6. Click Connect.

7. Complete the HighLevel OAuth authorization process.

8. Start a new Claude conversation after the connection is established.

Connect Codex

Codex supports the OpenAI-compatible MCP endpoint through its CLI or supported IDE extension.

To connect using the Codex IDE extension:

1. Open the Codex IDE extension.

2. Select the gear menu.

3. Select MCP servers.

4. Click Add server.

5. Choose Streamable HTTP.

6. Enter the following server URL:

   https://services.leadconnectorhq.com/mcp/openai/v2/

7. Select Authenticate.

8. Complete the HighLevel OAuth authorization process.

For Codex CLI, add the server using:

codex mcp add leadconnector --url https://services.leadconnectorhq.com/mcp/openai/v2/

Then begin OAuth authentication:

codex mcp login leadconnector

Follow the sign-in and authorization prompts.

Connect Claude Code

Claude Code supports the Anthropic-specific endpoint through its MCP configuration.

1. Open your terminal.

2. Add the hosted MCP server using:

   claude mcp add --transport http leadconnector https://services.leadconnectorhq.com/mcp/anthropic/v2

3. Complete the HighLevel OAuth authorization process when the browser opens.

4. Verify the connection using:

   claude mcp list

Connect Muse

Muse provides HighLevel MCP as an official app. Its connection uses the Muse-specific endpoint and OAuth authentication.

1. Open Muse.

2. Locate the app or MCP connection options.

3. Select the HighLevel app.

4. Begin the connection process.

5. Complete the HighLevel OAuth authorization flow.

6. Return to Muse and confirm that HighLevel is connected.

Authorize Your Connection with OAuth

OAuth allows your AI client to access approved HighLevel resources without requiring a shared account credential. During authorization, you select the available sub-account access and review the permissions requested by the connection.

1. Begin the OAuth authorization process from your AI client.

2. When the HighLevel authorization page opens, sign in with your HighLevel credentials.

3. Select the sub-account you want to connect.

4. If an eligible agency connection offers multi-sub-account access, select the appropriate sub-accounts.

5. Review the OAuth scopes and permissions requested by the integration.

6. Approve the authorization request.

7. Return to your AI client after authentication completes.

8. Confirm that the HighLevel MCP connection appears as connected.

Send Your First Project Management Request

Once your AI client is connected, you can describe the project work you want to manage. Start with a specific request that identifies the target List, the required action, and the relevant task details.

1. Open your connected AI client.

2. Start a new conversation or use an existing conversation with the HighLevel MCP connection available.

3. For clients that require manual tool selection, select the connected HighLevel MCP app or tool.

4. Enter a natural-language request.

   Example:

   "Create three tasks in the Onboarding List, due Friday, assigned to me."

5. Review the AI client's response and any operation results.

6. Verify the resulting tasks in HighLevel Project Management.

To verify the tasks in HighLevel:

1. Open HighLevel and switch to the relevant Sub-Account view.

2. Select Contacts from the left navigation menu.

3. Select the Tasks tab at the top of the page.

4. In the Spaces panel on the left, expand the relevant Space.

5. Select the Onboarding List.

6. Review the tasks and their assigned details.

Manage Project Work Using Natural Language

Once connected, your AI assistant can help with supported Project Management activities. You can describe the work you want to complete without manually navigating through every project action.

The following examples illustrate the supported capabilities identified in this release. Actual execution depends on the operations available to your connection and the permissions you approve.

Create and Update Tasks, Subtasks, Spaces, and Lists

Project Management organizes work through Spaces, optional Folders, Lists, tasks, and subtasks. MCP allows compatible AI clients to create and manage supported project objects using natural-language instructions.

You can request actions such as:

• Creating a new Space for a team or work stream.

• Creating a List inside a Space.

• Creating tasks in a selected List.

• Updating existing tasks.

• Creating or updating subtasks under a parent task.

Example request:

"Create a new Space for Client Onboarding and add an Onboarding List."

Work with Custom Statuses and Fields

Custom statuses help teams organize tasks by workflow stage, while custom fields store additional information relevant to a project.

Project Management MCP supports creating tasks with appropriate statuses and custom field values. This helps reduce the manual adjustments required after creating a task.

Example request:

"Create a task in the Onboarding List with the status and custom field values I specify."

For accurate results, reference the appropriate status and custom field values when describing the task.

Read and Post Task Comments

Task comments allow team members to discuss project work and keep important context alongside the relevant task.

With Project Management MCP, you can ask your AI client to read existing task comments or post a new comment. This lets you review discussions and contribute updates without opening the task interface.

Example requests:

• "Read the comments on the client onboarding task."

• "Add a comment to the client onboarding task with the latest project update."

Frequently Asked Questions

Q: Do I need to create a Private Integration Token to connect Project Management MCP?

No. The dedicated OAuth endpoints documented for Claude, ChatGPT, Codex, and Muse use a sign-in and authorization flow. These connections do not require you to manually create a Private Integration Token.

The older general-purpose MCP setup guide also documents token-based connections. Follow the authentication method appropriate for your selected endpoint.

Q: Can an agency connect multiple HighLevel sub-accounts to the same AI client?

Agency multi-sub-account connections are supported where the applicable installation option is available. During authorization, an agency can select the sub-accounts the connection may access.

Every request operates against one authorized sub-account at a time. Multi-sub-account availability is being rolled out and may not appear for every installation.

Q: Can I revoke an AI client's access to HighLevel after connecting it?

Yes. HighLevel's OAuth MCP documentation supports reviewing and revoking authorized connection access.

Revoking authorization removes the permissions previously granted to that connection.

Related Articles

• Introduction to Project Management

• Organize work with Spaces, Folders, and Lists

• Create a task or subtask

• Connect ChatGPT and Codex to LeadConnector MCP

• HighLevel MCP Multi-Account Support for Claude

• Connect HighLevel MCP to Muse