Integrations > BYO AI (Bring Your Own AI)
Connect your AI to Tallyfy
One-time setup for any AI platform
Section titled “One-time setup for any AI platform”You only need to do this once. Add Tallyfy’s MCP server URL to your AI platform, authorize with your Tallyfy account, and you’re connected - your AI can then read and act on your workflows through natural conversation.
What you’ll accomplish
Section titled “What you’ll accomplish”- Connect your AI to Tallyfy with one server URL
- Verify the connection works with a simple test prompt
- Understand where credentials are stored and what permissions you’re granting
Try it now
Section titled “Try it now”Setup
Open Claude.ai ↗ and go to Customize > Connectors, click + Add, then Add custom connector. On a Team or Enterprise plan, an Owner adds it under Organization settings > Connectors, and each member then clicks Connect next to it. Enter the name Tallyfy and the URL https://mcp.tallyfy.com. Save, then authorize with your Tallyfy account in the browser window that opens.
For detailed steps including Claude Desktop setup, see the Claude integration guide.
Prompt to try
What are my open tasks in Tallyfy?What happens
Claude calls the get_my_tasks tool and returns your current task list with deadlines and assignments. If you see a list of tasks, the connection is working.
Setup
In ChatGPT on the web, go to Plugins, select the plus button, then Add custom MCP server. You don’t need Developer Mode. Enter the name Tallyfy and the URL https://mcp.tallyfy.com, choose OAuth, and authorize with your Tallyfy account when prompted. Older screens start from Settings > Apps > Create instead (OpenAI called it Connectors before late 2025).
You need a ChatGPT Business, Enterprise, or Edu workspace for ChatGPT to make changes, and on Business only admins and owners can add Tallyfy. Pro accounts can connect with read access only.
For detailed steps, see the ChatGPT integration guide.
Prompt to try
What are my open tasks in Tallyfy?What happens
ChatGPT discovers the available Tallyfy tools automatically through the MCP server, then calls get_my_tasks and returns your task list. You’ll see a tool call step in the response before the results appear.
Setup
Copilot connects through a standard agent you build once in Copilot Studio: on the home page, select Other ways to build and create the agent. On the agent’s Tools page, select Add a tool > New tool > Model Context Protocol. Fill in a server name, a short description and https://mcp.tallyfy.com as the server URL, then choose OAuth 2.0 > Dynamic discovery. Create a new connection, sign in with your Tallyfy account, and add the tool to the agent. Then publish the agent to Teams and Microsoft Copilot.
For detailed steps, see the Copilot Studio setup guide.
Prompt to try
What are my open tasks in Tallyfy?What happens
Your Copilot agent calls Tallyfy’s get_my_tasks tool and shows your current task list in Copilot chat or Teams.
Setup
In the Gemini web app at gemini.google.com ↗[1], click Settings, then Connected Apps (it sits under Personal Intelligence if you don’t see it). Under Custom apps, click Add a custom app, enter https://mcp.tallyfy.com and click Next. Sign in with your Tallyfy account when asked. Gemini registers itself with Tallyfy automatically, so you can leave Advanced features alone. In a chat, type @ and pick Tallyfy so Gemini uses it.
Google offers this only on a personal Google account, for people 18 or over in the US, in English, with Keep Activity turned on. You add Tallyfy in the web app, and it then works in the Gemini mobile app too. These rules come from Google’s guide to custom apps in Gemini ↗[2].
Work and school accounts can’t add custom apps yet. For a team, a Google Cloud admin connects Tallyfy through Gemini Enterprise, and developers can use Gemini CLI or Gemini Code Assist. The Gemini setup guide covers all three.
Prompt to try
What are my open tasks in Tallyfy?What happens
Gemini connects to Tallyfy’s MCP server and retrieves your task list. The response includes task names, deadlines, and the processes they belong to.
How it works behind the scenes
Section titled “How it works behind the scenes”Your AI platform connects to Tallyfy’s MCP server using OAuth 2.1 with PKCE1. You sign in once with your Tallyfy credentials and authorize access - that’s it. Your AI platform stores the resulting token securely (in the OS keychain for desktop apps, in a platform vault for web apps), so you won’t need to authorize again unless you revoke access. From there, the MCP server exposes Tallyfy actions as named tools your AI can call, covering tasks, processes, templates, form fields, kickoff fields, automations, users, and more.
What permissions are granted
Section titled “What permissions are granted”When you authorize, you grant the following OAuth scopes2:
| Scope | What it covers |
|---|---|
mcp.tasks.read | View tasks and assignments |
mcp.tasks.write | Create, update, and complete tasks |
mcp.processes.read | View workflow runs |
mcp.processes.write | Launch, update, and archive processes |
mcp.templates.read | View templates and steps |
mcp.templates.write | Edit templates, steps, and automations |
mcp.forms.read | View form fields |
mcp.forms.write | Add and modify form fields |
mcp.users.read | View organization members and guests |
mcp.users.write | Invite and manage users |
mcp.automation.read | View automations |
mcp.automation.write | Create and modify automations |
You can revoke access at any time. In Tallyfy, go to your account settings and remove the authorized application. In your AI platform, disconnect the connector or integration from its settings panel.
Tips for better results
Section titled “Tips for better results”- Start with read-only questions before trying write operations. Get a feel for what your AI can see before asking it to make changes.
- Be specific - name the template, process, or person explicitly. “What tasks are assigned to Sarah on the Client Onboarding process?” gets a better result than “what’s Sarah doing?”.
- Your AI will ask for confirmation before making changes to Tallyfy data. That’s intentional - review the proposed action before approving it.
Take it further
Section titled “Take it further”Once you’re connected, here are some things worth trying next:
- Get a daily briefing of your tasks
- Find anything across your workflows
- Launch processes with pre-filled data