Task Hub — MCP-Driven Project Management

Overview

Task Hub is a modern, minimalist project management app I built to manage my own work. The core idea is to add an MCP server layer on top of the backend so I can feed PRDs and user stories to an AI agent and have tasks created automatically inside the app.

Problem

I needed a focused workspace to manage projects without bloated features. I also wanted a direct workflow from planning documents into structured tasks, without manual copy-paste.

Goals

• Build a clean, minimal project management experience for daily use.
• Create an MCP layer so AI agents can generate tasks from PRDs and user stories.
• Keep the system modular and extensible for future products.

Role and Responsibilities

I owned end-to-end product design and development: UX/UI, frontend implementation, backend services, database schema, and MCP server integration.

Constraints

• Solo build with limited time.
• Needed to keep performance fast and UX simple.
• Required a stable backend with future automation in mind.

Research and Insights

I benchmarked tools like Trello-style boards and modern SaaS dashboards. The key insight was that the UI must stay calm and minimal while the backend supports automation and integrations.

Information Architecture

• Landing page with product story and sections.
• Authenticated app with dashboards, boards, tasks, and settings.
• Backend modules for workspaces, boards, lists, tasks, and activities.
• MCP endpoint as the AI integration layer.

Design Approach

Modern, professional, minimalist, with subtle color accents. Clean typography, soft neutrals, strong hierarchy, and consistent spacing. The UI prioritizes clarity and focus over visual noise.

Core Features

• Boards, lists, and tasks with a clean Trello-like workflow.
• Activity tracking and task metadata for accountability.
• MCP server endpoint that allows AI agents to create tasks from PRDs and user stories.

Technical Highlights

• Stack: Bun, Elysia, Better Auth, Drizzle ORM, Postgres, Vite.
• MCP integration to expose task creation as tools for AI agents.
• Modular services and route groups for scaling.

Iterations

• Refined the visual system to remove heavy gradients and reduce noise.
• Simplified landing sections and tightened spacing.
• Adjusted tokens and components to support consistent theming.

Results

The app now serves as my personal project management hub and a testbed for MCP-driven automation.

Learnings

Minimal UI requires stronger information hierarchy. MCP workflows need clear schemas and predictable endpoints.

Next Steps

• Expand MCP tools for board and list creation.
• Add advanced automation rules.
• Improve reporting and timeline views.