Backlog.md

Backlog.md is a Markdown-native task manager and Kanban board for Git repositories that facilitates project collaboration between humans and AI agents.

Backlog.md is a task management tool that turns any Git repository folder into a project board using plain Markdown files. It organizes project collaboration between humans and AI agents through specific review checkpoints: reviewing the task spec, reviewing the implementation plan, and reviewing the generated code. The tool stores tasks as individual .md files directly within the repository, ensuring they remain as a permanent, human-readable record. It includes a terminal-based Kanban board and a local Web UI for managing tasks with drag-and-drop support. It integrates with AI assistants via the Model Context Protocol or standard CLI instructions, supporting tools like Claude Code, Gemini CLI, Codex, and Kiro.

Backlog.md at a glance

Best for: Software engineer, Product manager
Runs on: macOS, Linux, Windows, Web
Pricing: MIT-licensed and open-source
Released: June 2025
Category: AI coding tool
Website: backlog.md

Key features

Markdown-native task tracking in Git
Terminal-based Kanban visualization
Local Web UI for drag-and-drop task management
Three-step AI review workflow for specs, plans, and code
Acceptance criteria and Definition of Done checklists
Milestones and task dependency tracking
Fuzzy search for tasks and decisions

Use it for

Managing AI-assisted coding projects
Reviewing task specifications and implementation plans
Visualizing project progress in a terminal Kanban board
Tracking task history and decisions within a Git repository

Alternatives to Backlog.md

Better Backlog — Filters GitHub issues locally using natural language and custom criteria.
DotMD — Reviews documents like code with line-specific agent fixes.
AGENTS.md — It separates agent-specific instructions from human-focused README files.
docmd — Compiles Markdown into websites and AI agent context formats with one command.
Mdt — Avoids unnecessary rereading, token use, and distraction for agents.
jevgrep — Reduces coding agent costs by providing targeted, relevant repository context.
Markdoc — Real-time collaboration with source and preview side by side.
GitBrain — Intelligently splits large code changes into multiple customizable AI commits.
Atlas — Centralize agent control, session logging, and project git state in one workspace.
Enzyme — It indexes markdown notes locally so AI coding agents can cite them.
Crit — It keeps human feedback localized and precise for AI agent code iterations.
Ret — Manage multiple AI coding agents across projects from one unified desktop interface.

All Backlog.md alternatives