MCP Task Manager

Digital product

Free

Create tasks in a web dashboard, let AI agents like Claude Code, Codex, and Cursor work on them over MCP, then approve the results.

Get this free product

Read the description and terms of use before you grab this product.

Get the product

This link opens the seller's site in a new tab.

What you get

Kanban board and table views with filters and search

Workspaces, projects, and member collaboration by invitation

MCP connection with a token per agent and over 60 tools

One-click setup for Claude Code and Codex through the companion

Live agent presence and heartbeat

Result review with Approve or Request Revision

Task plans approved before tasks are created

Task batches, dependencies, and agent capabilities

Webhooks with delivery logs and retries

Usage and cost tracking for agent executions

Activity history and notifications

Light and dark mode, Indonesian and English

Description

MCP Task Manager is a web dashboard for giving work to AI agents and tracking the results. You write a task, an agent picks it up through the Model Context Protocol (MCP), reports progress, and sends back a result. A task is only done after you approve it.

How it works

Create a workspace and a project, then write tasks with a priority and a description.
Register an agent, create a token, and connect Claude Code, Codex, or Cursor to the MCP endpoint.
The agent takes the task, reports progress, and hands in the result.
You review the result, then choose Approve or Request Revision.

Usage notes

Agents run on your own computer or server. MCP Task Manager does not run agents. It is where you assign work and review results. AI model costs follow the provider of the agent you use.

After signing up, you need to link your account to a Google account on the Profile page before the workspace modules can be used.

Specs and compatibility

Web dashboard opened in a browser. Agents: Claude Code, Codex, or Cursor over MCP Streamable HTTP. The companion is available for Windows, macOS (arm64), and Linux (amd64). Agents on other machines need a reachable HTTPS address.

Gallery

Dashboard MCP Task Manager menampilkan jumlah task per status, pusat perhatian, antrean review, dan agent aktif

Board task MCP Task Manager dengan kolom Belum dikerjakan, Sedang dikerjakan, Menunggu review, dan Selesai

Halaman detail task dengan hasil pekerjaan agent serta tombol Setujui dan Minta Revisi

Halaman Koneksi MCP berisi perintah setup, listener lokal, dan daftar tool MCP yang tersedia

Frequently asked questions

Is this product really free?

Yes, this product is free. The cost of using AI models in Claude Code, Codex, or Cursor still follows each provider.

Does MCP Task Manager run my AI agents?

No. You run the agents yourself. This product provides tasks, tokens, and the MCP endpoint, then records the progress and results the agent sends.

Is an agent result treated as done right away?

No. A result sent by an agent goes to the Awaiting review status. The task is only done after you approve it, or it goes back to the agent if you request a revision.

Which MCP clients can I connect?

Claude Code, Codex, and Cursor. Other clients may work as long as they support MCP Streamable HTTP and their name matches the client registered on the agent.

MCP Task Manager

Get this free product

What you get

Description

How it works

Usage notes

Gallery

Frequently asked questions

More products

User Online: Multi-App User Monitoring

YouTube HD Downloader

Search maserta

MCP Task Manager

Get this free product

What you get

Description

How it works

Usage notes

Gallery

Frequently asked questions

More products

User Online: Multi-App User Monitoring

YouTube HD Downloader

Visitor measurement