About TaskFlow

What is TaskFlow?

TaskFlow is a local-first task and time tracking application designed specifically for AI-powered development workflows. It provides 50 MCP (Model Context Protocol) tools that let AI agents autonomously manage tasks, track time, log debug breadcrumbs, and coordinate with other agents.

Key Principles

• Local-first: All data stays on your machine in a SQLite database. No cloud accounts, no telemetry, no data leaving your system.
• Agent-native: Built from the ground up for AI agents. Every feature is accessible via MCP tools, not just a human GUI.
• Real-time sync: SSE (Server-Sent Events) broadcasts changes instantly from the MCP server to the Tauri desktop app.
• Multi-agent: Multiple AI agents can collaborate on the same project, communicate via messages, and coordinate through task dependencies.

Architecture

1AI Agent (Claude / Codex / Cursor)
2        ↓ stdio (MCP protocol)
3taskflow-mcp (Node.js server)
4        ↓                ↓
5   SQLite DB        SSE :3456
6                        ↓
7                  Tauri Desktop App
8                  (React + Dexie.js)

Tech Stack

Features at a Glance

• 50 MCP Tools: Tasks, projects, timers, analytics, notifications, agent messaging, terminal capture
• Dependency Graph: Interactive DAG visualization with automatic cycle detection
• Activity Pulse: Real-time event log showing agent actions, debug entries, and status changes
• Agent Inbox: Remote Q&A system where agents ask questions and users respond from any device
• Multi-Agent Collaboration: Agent registry, messaging, broadcasts, and coordinated task assignment
• Remote Control: Optional relay server for controlling agents from anywhere (outbound-only, no VPN needed)

License

TaskFlow is open source under Apache 2.0 with Commons Clause: free to use, modify, and share, but not to sell as a product.

• GitHub Repository
• npm Package