Taskworld MCP. Manage projects and tasks in conversation.

Listen up. This server connects your AI agent right into Taskworld, giving it the muscle to handle projects and tasks—all from plain language prompts. You don't gotta open a web browser or click through menus; you just talk to your agent, and it gets the job done in your workspace.

To get started, your agent first needs credentials. It runs login to grab an access token for Taskworld, making sure everything else works smoothly after that initial handshake. Once authenticated, your agent can map out what's available: you can use list_workspaces to see every workspace your account has access to.

If you need the specifics on one location, calling get_workspace pulls all the detailed information for that single Taskworld workspace.

When it comes to organizing work, things get real. Your agent handles project lifecycle management. You can kick off a whole new effort by running create_project. Need to change something about an existing project—maybe rename it or mark the whole thing as finished? The update_project tool lets you manage those details easily.

Meanwhile, your agent can list all the projects inside any workspace using list_projects, giving you a full rundown of what's running.

Managing the actual tasks is where this server shines. First, to get oversight, you can run list_tasks against a specific project ID to see every task that’s currently active on it. If a new job pops up, your agent can generate and add a brand new item using create_task. You don't just list tasks; you manage them.

You can use update_task to change critical properties—like shifting the due date, reassigning the task owner, or tweaking the title itself. When a job is done, your agent marks it complete by calling complete_task, updating its status instantly.

For keeping tabs on progress and history, you've got two tools. You can use list_comments to grab every single message posted on either a task or an entire project thread, giving you the whole conversation history right away. If someone needs to talk about it—say, they found a bug or need clarification—your agent posts that comment directly into the flow using create_comment.

This means all communication stays exactly where it belongs: attached to the specific task or project.

Basically, your AI client uses this server to handle everything from initial setup and discovery (list_workspaces, get_workspace) to building out structure (create_project), assigning work (create_task, update_task), tracking progress (complete_task), and keeping the communication loop closed by posting or retrieving messages (create_comment, list_comments). You'll be able to manage your whole operational pipeline without ever leaving your chat window.