Connect Taskonio to the agent you already use.

Taskonio MCP connects compatible AI clients to your tasks, team spaces, and sprints. Ask your agent to find the work, help plan it, or make the changes you authorize.

Choose your client

Less pasting context. More working with it.

Find the next thing to do

Ask for the tasks assigned to you in the current sprint, or paste a task URL to inspect its description and discussion before starting.

Turn context into a plan

Review a backlog, draft a specification, or propose a task breakdown. Be explicit about whether you want suggestions or actual changes.

Keep the team’s work current

Ask the agent to update an assignee, add a comment, or move selected tasks into a sprint. Verify the result in Taskonio.

Quick setup

Pick your client and connect.

Connect to Taskonio's production MCP endpoint, then sign in and choose the organization your agent can access. Most clients guide you through authorization automatically.

Connection details

Transport
Streamable HTTP
Authentication
OAuth 2.1 with PKCE

Claude Code

Run the command, then open /mcp in Claude Code to authenticate the server. Sign in to Taskonio and select the organization you want to use.

claude mcp add --transport http taskonio https://api.taskonio.com/v1/mcp
Claude Code MCP documentation

Claude Desktop

Add Taskonio as a custom remote connector in Claude’s Connectors settings, using the URL below. Complete the authorization flow. On a managed team account, an owner may need to add the connector first. This hosted server uses the remote-connector flow, not a local desktop-extension configuration.

https://api.taskonio.com/v1/mcp
Claude remote connector instructions

Cursor

  1. Open Cursor’s MCP configuration, or create .cursor/mcp.json in your project.
  2. Add the taskonio entry below to the existing mcpServers object without removing other servers.
  3. Save the configuration and complete authorization when prompted.
  4. Check that the Taskonio tools are enabled before asking the agent to use them.
{
  "mcpServers": {
    "taskonio": {
      "url": "https://api.taskonio.com/v1/mcp"
    }
  }
}
Cursor MCP documentation

Windsurf

Open the MCP configuration from Cascade’s MCP controls and add the server below to mcp_config.json. Keep existing entries, save, and authorize the connection. Use the configuration location shown by your installed client version.

{
  "mcpServers": {
    "taskonio": {
      "serverUrl": "https://api.taskonio.com/v1/mcp"
    }
  }
}
Cascade MCP documentation

Other MCP clients

Use Streamable HTTP, connect to the endpoint below, and complete the OAuth flow opened by your client.

https://api.taskonio.com/v1/mcp

Authentication

Secure access, handled by your client.

Taskonio uses OAuth 2.1 with PKCE. You choose the organization the client may access.

  1. Connect

    Your client initiates a connection to the MCP endpoint.

  2. Sign in

    Taskonio opens so you can authenticate.

  3. Choose access

    Select the organization you want to authorize.

  4. Continue working

    The client receives access and handles token refresh.

OAuth discovery endpoints

GET https://api.taskonio.com/.well-known/oauth-authorization-server
GET https://api.taskonio.com/.well-known/oauth-protected-resource
Access token
1 hour

Refreshed automatically

Refresh token
30 days

Available tools

Give your agent useful access to real work.

The MCP exposes focused tools for task operations, workspace context, organization data, planning, and Taskonio AI.

Tasks

create_task
Create a new task in a workspace
get_task
Get task details with comments and labels
update_task
Update status, assignee, priority, and other properties
delete_task
Permanently delete a task
search_tasks
List and search tasks with pagination
add_comment
Add a comment to a task
get_subtasks
Get subtasks or tasks under an epic
resolve_task_url
Fetch task details from a Taskonio URL

Workspaces

list_workspaces
List every workspace you can access
get_workspace
Get workspace statuses, labels, and views
create_sprint
Create a new sprint
list_epics
List epics in a workspace
list_workspace_members
List members of a workspace

Organization

list_organizations
List your organizations
get_organization_members
List members in an organization
get_notifications
Get your Taskonio notifications

Planning

get_planning_session
Get the latest active planning session
create_planning_session
Start a new planning session
move_tasks_to_sprint
Assign tasks to a sprint

AI-powered

ai_suggest_description
Generate a task description from a title
ai_breakdown_task
Break a task into subtasks

Usage examples

Ask in plain language.

Once connected, use direct requests or paste a Taskonio task URL into your conversation.

> Show me all tasks in the Application workspace

> Create a bug report: "Login page crashes on Safari 17"

> What is the status of task #42?

> Break down task #48 into subtasks

> Move tasks #101, #102, and #103 to the current sprint

> Assign task #55 to me and set priority to High

> Analyse this task: https://app.taskonio.com/org123/space/abc/task/48

Troubleshooting

Common connection fixes.

"Authentication failed" or "Unauthorized"+
  • Make sure you are signed in to Taskonio.
  • Remove and add the MCP server again.
  • Check that your account can access the organization you selected.
"Connection refused" or timeout+
  • Verify that the URL is https://api.taskonio.com/v1/mcp.
  • Check your internet connection.
  • If the server is temporarily unavailable, wait a few minutes and retry.
Tools are not showing+
  • Restart your AI client after adding the MCP server.
  • Check that authorization completed successfully.
  • Some clients require MCP tools to be enabled in settings.
Session expired+
  • Sessions expire after 30 minutes of inactivity.
  • Your client should reconnect automatically. Restart it if it does not.

Team configuration

Share the connection with your project.

For Claude Code, add this configuration to .mcp.json in the project root. Each teammate approves and authenticates the connection separately. Cursor uses .cursor/mcp.json with the Cursor configuration above. Share server settings, never private tokens.

{
  "mcpServers": {
    "taskonio": {
      "type": "http",
      "url": "https://api.taskonio.com/v1/mcp"
    }
  }
}

Questions

Know what you’re connecting.

What is Taskonio MCP?

Taskonio MCP is a hosted Model Context Protocol server. It lets compatible AI clients use Taskonio tools for tasks, team spaces, organizations, sprint planning, and AI-assisted specifications and breakdowns.

Which endpoint should I use?

Use https://api.taskonio.com/v1/mcp with Streamable HTTP. Sign in to Taskonio and select the organization when your client asks you to authorize the connection.

Can my agent change tasks, or only read them?

It can do both. Available tools include creating and updating tasks, adding comments, moving work into sprints, and deleting tasks. Review your client’s tool-approval settings before using write or delete tools, and be explicit about the changes you want.

Does every MCP change go through Decisions?

No. An authorized MCP client can create and update tasks directly through its tools. That is separate from reviewing Vibe drafts or recommendations in Decisions. Use the approval controls in your AI client to manage what the agent can do.

Does connecting MCP give my whole team the same access?

No. Sharing a server URL or project configuration is not sharing an authenticated session. Each teammate must complete authorization with their own account. Never commit access tokens or private credentials to a shared configuration.

Do I need Slack or GitHub connected to use MCP?

No. Taskonio MCP is a separate connection for an AI client. Slack provides personal Action and assignment notifications plus task-link previews, while GitHub links development activity. You can use the connection that fits your workflow.

Put Taskonio in the workflow you already use.

Create your workspace, then connect the agent that helps you plan and ship.

Start free