THE PROBLEM

Using AI to code, but manually managing your board.

Your AI finishes a bug fix, but you still have to open Jira to update status.
Context switching between terminal, editor, and browser kills flow state.
Cloud-based trackers index all your private project details.

THE SOLUTION

Autonomous task management via MCP.

Provide 31 tools to Claude/Cursor. The agent fixes a bug, creates a task, and marks it done.
Zero context switching. You watch the board update via real-time SSE while you review code.
100% local SQLite database. Your data never leaves your machine.

Watch an AI Agent Manage a Project

A Claude Code agent implements a feature from scratch — creating tasks, tracking time, logging progress, and marking work done. No human touched the board.

Built for the Agentic Era

Everything you need to manage complex software projects, designed specifically to be driven by LLMs via the Model Context Protocol.

31 MCP Tools for AI Agents

Install the server with one command. Your AI gets tools for tasks, projects, timers, and analytics. It creates tasks, logs breadcrumbs, and marks work done autonomously.

~/.claude.json

"mcpServers" : {
   "taskflow" : {
     "command" : "taskflow-mcp" ,
     "args" : []
  }
}

Real-Time Live Sync

Agent writes to SQLite → SSE fires → Tauri app updates IndexedDB → UI re-renders instantly.

AI

Visual Task Dependencies

Tasks have dependencies. Dependencies have a graph. Interactive DAG visualization with automatic cycle detection. Your AI agent can query the graph to understand blockers before starting work.

TaskFlow Dependency Graph — interactive DAG with project nodes, status colors, and task connections

Concurrent Timers

Track time across multiple tasks. Every session is logged with precise timestamps.

01:12:45

PAUSED

Activity Pulse

AI agents use the `log_debug` tool to leave markdown breadcrumbs—hypotheses, errors, findings.

10:42 AM • AI Agent

Logged debug: Found null pointer in parser.

10:45 AM • AI Agent

Marked task as done.

Built-in Terminal

Ctrl+K opens a full xterm.js terminal. Create tasks, navigate, and manage state without a mouse.

taskflow> task create "Fix CORS"
Created task #42
taskflow> timer start 42 _

See It in Action

Real screenshots from the TaskFlow desktop app. Dark theme, zero fluff.

System Performance — Analytics

TaskFlow Analytics — status distribution, time per project, focus metrics

Built-in Terminal — xterm.js

TaskFlow Terminal — Ctrl+K command palette with task list and detail view

Activity Pulse — Event Log

TaskFlow Activity Pulse — timer events, status changes, debug logs from AI agents

Task Detail — Markdown + Sessions

TaskFlow Task Detail — markdown description, session timeline, dependencies, tags

Floating Timer Bar

TaskFlow Floating Timer — play, pause, stop with elapsed time

Neon Flux Core — Settings

TaskFlow Settings — SSE port, notification engine, status colors, dev tools

Archived Success — Completed Tasks

TaskFlow Archive — completed tasks across all projects with time tracking

Get Started in 3 Steps

1

Download TaskFlow Desktop

Choose your Linux distribution format (macOS/Windows planned):

AppImage

Universal — run anywhere

2

Install the MCP Globally

Make the TaskFlow MCP server available to your system via npm.

Terminal

# Install the package globally
npm install -g @dalmasonto/taskflow-mcp

3

Configure Your Agent

Create a .mcp.json file in your project root (or home directory for global access). Claude Code will auto-detect it.

.mcp.json (project root or ~/.mcp.json)

{
   "mcpServers" : {
     "taskflow" : {
       "command" : "taskflow-mcp" ,
       "args" : []
    }
  }
}

Grant permissions

// ~/.claude/settings.json or
// .claude/settings.local.json (per project)

{
   "permissions" : {
     "allow" : [
       "mcp__taskflow__*"
    ]
  },
   "enableAllProjectMcpServers" : true ,
   "enabledMcpjsonServers" : [
     "taskflow"
  ]
}

The agent calls get_agent_instructions on startup — it learns how to use TaskFlow automatically. No prompt engineering needed.

Local-First Architecture

Designed for privacy and speed. No cloud sync, no accounts. Just an SQLite file and an SSE broadcast.

Claude Code / Cursor MCP Client

stdio

taskflow-mcp (Node.js) Server

FS

SQLite DB

SSE :3456

Tauri App

React 19 Tauri v2 Dexie.js better-sqlite3

Optimized for AI Context

MCP responses are condensed. Null suppression and compact JSON mean your AI has more context window available for actual reasoning, not bloated payloads.

61%

Token reduction vs v1.0.0

72%

Smaller list_tasks payload

list_tasks (v1.0.4) ~100 chars

{
   "id" : 42 ,
   "title" : "Build dependency graph" ,
   "status" : "done" ,
   "dependencies" : [ 38 , 39 ]
}

MCP Tools Reference

31 tools across 8 domains. Agent instructions are built-in — your AI learns TaskFlow automatically.

Tasks

create_task, list_tasks, get_task, update_task, update_task_status, delete_task, bulk_create_tasks, search_tasks

8

Projects

create_project, list_projects, get_project, update_project, delete_project, search_projects

6

Timers

start_timer, pause_timer, stop_timer, list_sessions

4

Activity

get_activity_log, log_debug, clear_activity_log

3

System

get_analytics, get_timeline, list_notifications, mark_notification_read, mark_all_notifications_read, clear_notifications, get_setting, update_setting, get_agent_instructions, clear_data

10

Your AI Agent's
Task Manager

Using AI to code, but manually managing your board.

Autonomous task management via MCP.

Watch an AI Agent Manage a Project