Sally MCP. Manage tasks and projects without leaving your chat.

Sally connects your project management system right into your AI client. You'll manage projects and track team work without jumping between apps or clicking through a dozen menus. It gives your agent direct access to the core data—the stuff you actually use every day.

Setting Up Shop

You can first run check_sally_health to confirm that the connection is solid and everything's running right on Sally's end. If you need to know who your AI client is speaking for, you just call get_profile; it pulls up your authenticated identity and confirms what permissions you've got in the workspace.

To see all the big picture stuff—every project currently active—you use list_projects. You can also dive deep into a specific job by running get_project with an ID, which gives you every detail about that single initiative.

Creating and Listing Projects

Need to start something new? Run create_project, and Sally initializes the whole record, giving it its own unique metadata. When you're done setting up projects, you can pull a list of tasks belonging only to one specific job using list_tasks. If you just want to see all outstanding tasks across every project—no filtering—list_tasks handles that too.

Managing Tasks: The Core Work

The heart of the system is managing individual tasks. You can generate a brand new task by calling create_task, and when you do, you get to set specific details right away: you can assign priority levels from P1 through P4, define statuses, and add labels that make sense for your team's process.

You never have to manually update something twice. If a status changes or someone needs to adjust the due date, you call update_task. This function lets your agent modify an existing task using any combination of fields; only the data you provide actually gets changed in Sally.

When the team is collaborating on a piece of work, they don't write emails. They just use add_comment to post visible status notes or comments directly onto the task record. You can get all the full details for one specific item by running get_task with its ID. If you need a summary of every single project task across the board, you run get_board, and it pulls up that comprehensive Kanban view, grouping tasks automatically by their current status column.

Tracking Time and Billing

For billing purposes, your agent can pull two kinds of timesheet reports: one for the entire workspace using get_timesheet_report or a specific report just for one project. These functions give you all the hours tracked and the data needed to bill clients.

It’s Simple: You don't need to write complex code; you tell your AI client what you want—like, 'Give me the P2 tasks for Project Alpha,' or, 'Update Task 789 status to Done.' Your agent handles making those specific calls (list_tasks, update_task) and bringing you back exactly what you asked for.

You get real-time project status, task changes, and billing reports all through simple conversation with your AI client.