A markdown-native task manager and Kanban visualizer designed for any Git repository, facilitating project collaboration between humans and AI agents.
One-minute overview#
Backlog.md turns any folder with a Git repo into a self-contained project board powered by plain Markdown files and a zero-config CLI. Whether you're an individual developer or a team, you can use it to manage project tasks while collaborating seamlessly with AI agents.
Core value: Complete local task management through Markdown files without requiring cloud services, while supporting integration with multiple AI tools.
Getting Started#
Installation difficulty: Low - Simple command-line installation gets you started quickly
# Installation command
npm i -g backlog.md
# or
bun add -g backlog.md
# or
brew install backlog-md
Is this suitable for me?
- ✅ Individual developers/small teams: Need a completely local, cloud-service-free task management tool
- ✅ AI-assisted development: Want to collaborate with AI agents like Claude, Gemini on projects
- ❌ Complex project management: If you need advanced features like Gantt charts, resource allocation, you might need more specialized project management software
- ❌ Large team collaboration: While it supports multiple users, it doesn't have the complex permissions and auditing features of enterprise project management software
Core Capabilities#
1. Markdown-native task management - Streamlined documentation#
- Each task is a standalone .md file stored in the
backlogfolder - File format:
task-<task-id> - <task-title>.mdfor easy identification and management Real value: Task documentation happens automatically without extra steps, seamlessly integrating with existing Git workflows
2. AI collaboration capabilities - Human-AI development partnership#
- Supports multiple AI agents including Claude Code, Gemini CLI, Codex
- Deep integration with task management through the Model Context Protocol (MCP) Real value: AI agents can directly understand, create, and execute tasks, enabling automated development workflows
3. Dual interface experience - Flexible working methods#
- Terminal board:
backlog boardcommand displays a live board in your terminal - Modern web interface:
backlog browserlaunches a responsive web UI for visual task management Real value: Choose the interface that best fits your environment and preferences, boosting productivity
4. Powerful search and querying - Quickly find information#
- Fuzzy search functionality across tasks, documents, and decisions
- Rich query commands supporting filtering by status, priority, assignee, and more Real value: Quickly locate needed information among many tasks, improving project discoverability
5. Task dependency management - Clear workflows#
- Supports task dependency setup with circular dependency prevention
- Visual execution sequences clearly showing relationships between tasks Real value: Properly prioritize and sequence tasks to avoid work conflicts
Tech Stack & Integration#
Development language: TypeScript/JavaScript Main dependencies: Built on Node.js, installable via npm, bun, or Homebrew Integration method: CLI tool + Web interface + MCP protocol (for AI integration)
Ecosystem & Extensions#
- Plugins/Extensions: Integrates with various AI agents through the MCP protocol including Claude Code, Gemini CLI, Codex
- Integration capabilities: Fully integrated with Git workflows, all data stored as Markdown files in the local repository
Maintenance Status#
- Development activity: Actively developed with support for multiple installation methods and continuous AI tool integration
- Recent updates: Includes complete MCP integration guide, showing developers are actively following AI tool developments
- Community response: Provides detailed documentation and examples, indicating focus on user experience
Commercial & Licensing#
License: MIT
- ✅ Commercial: Commercial use allowed
- ✅ Modification: Code modification allowed
- ⚠️ Restrictions: Must include original license notice and copyright information
Documentation & Learning Resources#
- Documentation quality: Comprehensive with detailed CLI reference and integration guides
- Official documentation: Detailed README and documentation included in the GitHub repository
- Example code: Complete command-line usage examples and web interface screenshots provided