Taskmaster is a task management system built specifically for AI-driven development workflows. Instead of manually tracking tasks in a separate tool, you describe your project in a PRD and Taskmaster breaks it into structured, sequenced tasks that your AI coding assistant can work through one by one.
The tool integrates directly into your AI editor via MCP (Model Context Protocol) or as a CLI. In Cursor, you can install it with one click. In Claude Code, a single command adds the MCP server. It also works with Windsurf, VS Code, Amazon Q, and Codex CLI. The AI assistant can then parse PRDs, list tasks, show the next task to work on, expand tasks into subtasks, and track progress automatically.
Taskmaster supports 12 AI providers including Anthropic, OpenAI, Google Gemini, Perplexity, xAI, OpenRouter, Mistral, Groq, and Ollama. You can configure different providers for main, research, and fallback model roles. The research command uses a dedicated research model to investigate best practices before implementation.
Key features include tagged task lists for multi-context management (per branch, per feature), a research command with project context awareness, loop command for automated task execution with streaming, git integration for branch-based task creation, and user-defined metadata fields. Tool loading is configurable: all 36 tools (~21K tokens), standard 15 tools (~10K tokens), or core 7 tools (~5K tokens) to manage context window usage.
With 26,900 GitHub stars and 72 contributors, Taskmaster is one of the most popular AI development workflow tools. The npm package (task-master-ai) has 92 releases and is actively maintained under MIT with Commons Clause license.