Connect Tooling Studio to AI apps with MCP
Tooling Studio MCP lets you connect your workspace to compatible AI apps so you can ask for help with Tasks and CRM in plain language. The AI can find the right workspace, board, list, contact, or organization for you, then act using your existing Tooling Studio access.
- Works with your existing Tooling Studio permissions
- Helps with both Tasks and CRM
- Designed for plain-language requests, not API work
This feature is still in private release. If you enable and use it now, do so at your own risk.
Ask your AI app for help with Tasks and CRM in normal language
Once connected, the AI can look up the right workspace, board, list, task, contact, or organization before it acts.
"What tasks are due today?"
"Move everything from Backlog to Doing that mentions onboarding."
"Find Olivia Bennett in CRM and add a note."
"Create a contact for Northstar Labs and place it in Prospects."
The AI usually starts by looking at the teams, boards, lists, pipelines, and records you can access so it can understand where your request belongs.
It finds the most likely task, contact, organization, board, or list based on the names and details in your request, even if you do not remember the exact wording.
Once it finds the match, it can read, create, update, move, assign, tag, comment, or organize the record depending on what you asked it to do.
Not publicly released yet. Request access first and use it with care while the feature is still evolving.
What MCP means here
A connection layer between Tooling Studio and your AI app
You connect Tooling Studio once inside an AI app that supports MCP. After that, you can talk to the AI normally. You do not need IDs, APIs, or a mental map of every board and list. The AI can look up the structure it needs first, then carry out the action you asked for.
Plain-language requests
Ask for help the way you would ask a teammate. MCP gives the AI the bridge it needs to search, read, create, update, move, and organize the right records.
Your permissions still apply
MCP does not grant extra admin powers. The AI app can only reach the teams, boards, lists, tasks, contacts, and organizations you already have permission to access.
Remote setup, not local setup
Tooling Studio MCP gives you a URL. In most AI apps, that means you want the remote connector or remote server flow, not a local command-line install.
Tasks and CRM
What MCP can do inside Tooling Studio
Tooling Studio MCP is meant to help normal users work faster across Tasks and CRM. It can look things up, ask follow-up questions when something is ambiguous, and take action using the access you already have.
What it can do in Tasks
- Create, rename, and reorganize task boards and lists
- Search tasks by title, description, board, list, assignee, or due date
- Create tasks, move them, complete them, or reopen them
- Update titles, descriptions, due dates, checklists, comments, assignees, and tags
- Link tasks to CRM contacts or organizations
"What tasks are due today?"
"Show me everything in the Product board that is due this week."
"Create a task called Prepare launch checklist in Backlog."
"Move Prepare launch checklist to Doing."
"Mark all onboarding tasks assigned to me as complete."
What it can do in CRM
- Search for contacts and organizations across your workspace
- Create contacts, organizations, boards, and pipeline lists
- Move records through a pipeline and assign owners
- Add notes, comments, tags, links, and related task connections
- Update profile details and place records in the right list or stage
"Find Olivia Bennett in CRM."
"Create a new contact for Olivia Bennett with [email protected]."
"Add Olivia Bennett to the Prospects list."
"Move Olivia Bennett from Prospects to Qualified."
"Create an organization called Northstar Labs with website northstarlabs.example."
What MCP can and cannot do
Use it for real work, but still use judgment
MCP can save a lot of clicks, but it is still a connection between an AI app and your workspace. It helps most when you stay specific and review important actions carefully.
What it can do well
- Work with the Tooling Studio data you already have access to
- Handle follow-up actions like moving, tagging, assigning, commenting, and linking
- Ask follow-up questions when your request is ambiguous
- Help you move faster across Tasks and CRM without memorizing structure
What it still cannot do
- Bypass your existing Tooling Studio permissions
- Magically know which record you mean if multiple things look the same
- Guarantee perfect results every time without your judgment
- Override provider-specific limits in the AI app you connect
How to connect an AI app to Tooling Studio MCP
The setup is meant to stay simple. Turn on MCP access in Tooling Studio, copy the MCP URL, add it to the AI app, review the sign-in flow, and then start with a simple request first.
Turn on MCP access in Tooling Studio
Once your workspace has access, open Settings -> Account -> AI, then turn on MCP access. This is the switch that allows external AI apps to connect at all.
Not yet publicly available
If you want to enable MCP for your workspace, request early access first.
Copy your Tooling Studio MCP URL
In the same AI settings area, Tooling Studio shows your MCP URL. Copy it exactly. This is the URL you will paste into the AI app when it asks for a connector or MCP server.
Add the URL in your AI app
Look for the AI app's MCP, connector, or remote server setup flow. Paste in your Tooling Studio MCP URL there. If the app offers both local and remote setup, choose the remote option.
Sign in with Google and review the request
Your AI app should open a sign-in and approval flow. Use the same Google account you already use with Tooling Studio, then confirm the app details before approving the connection.
Start with a simple request first
Begin with something easy like checking due tasks, finding a contact, or listing a pipeline. Once you trust the setup, move on to edits and updates.
Before you connect
- This is a private early-access feature and is not publicly released yet.
- Use an AI app that supports MCP connectors or remote MCP servers.
- Sign in with the same Google account you already use with Tooling Studio.
- Only connect Tooling Studio inside AI apps and workspaces you trust.
Managing access and disconnecting later
You stay in control of the connection. In Tooling Studio, go back to Settings -> Account -> AI at any time to review connected MCP apps, disconnect one app, or turn MCP access off entirely. Disconnecting an app removes its access right away. Turning MCP access off blocks all MCP connections until you enable it again.
If you want to fully clean up, remove the connector from the AI app as well after disconnecting it in Tooling Studio.
Official setup guides
Use the guide that matches your AI app
Provider support changes quickly. Check the latest official docs before you connect. Tooling Studio MCP is a remote URL-based setup, so prefer guides that explain remote connectors or remote MCP servers.
ChatGPT
OpenAI documents ChatGPT developer mode and MCP support in its official platform docs. This is the right starting point if your team is connecting Tooling Studio through ChatGPT, but exact availability depends on plan and workspace setup.
Claude
Anthropic documents custom connectors using remote MCP for Claude and Claude Desktop. This is the clearest guide if you want a connector-style setup flow.
Cursor
Cursor documents MCP for both remote and local setups. Use the remote server path when you are pasting in your Tooling Studio MCP URL.
Official MCP docs
If you want the big-picture explanation of hosts, clients, servers, tools, resources, and prompts, start with the official MCP architecture overview.
Frequently asked questions
Answers to the most common questions about Tooling Studio MCP, early access, setup, permissions, and disconnecting later.