Backlog.md
Backlog.md - A tool for managing project collaboration between humans and AI Agents in a git ecosystem
Backlog.md is an MCP server and CLI tool that provides AI agents with structured task management capabilities through the Model Context Protocol. It enables spec-driven AI development by giving agents the ability to create, read, and manage project tasks stored as Markdown files in Git repositories. The server exposes task management operations, workflow instructions, and project context to AI coding assistants, helping teams review AI-generated code at three checkpoints: specification review, implementation planning, and code review. Installation requires running 'backlog init' in a project, then connecting via MCP clients like Claude Code or Codex.
Key Features
Use Cases
- 01Enabling AI agents to decompose features into reviewable tasks with acceptance criteria before writing code
- 02Structuring AI coding sessions around single-task context windows to keep code reviews manageable
- 03Creating a permanent Git-based record of AI agent work with task descriptions, plans, and completion notes
- 04Managing human-AI collaborative development workflows with clear handoff points for review and approval
- 05Running project task management entirely from the terminal or browser without external SaaS dependencies
- 06Tracking AI-generated implementation plans and verifying approaches before committing to code changes
Related MCP Servers
View moremarkitdown
Python tool for converting files and office documents to Markdown.
firecrawl
The context API to search, scrape, and interact with the web at scale. 🔥
prompts.chat
f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.
langflow
Langflow is a powerful tool for building and deploying AI-powered agents and workflows.
Backlog.md — FAQ
What is the Backlog.md MCP server?+
Backlog.md is an MCP server that provides AI coding assistants with task management capabilities, enabling agents to create, read, and manage project tasks stored as Markdown files. It implements spec-driven AI development with three review checkpoints to keep AI-generated code manageable and reviewable.
How do I install and connect Backlog.md to my AI client?+
Install globally with 'npm i -g backlog.md' or 'bun add -g backlog.md', run 'backlog init' in your project, then add the MCP server using your client's command (e.g., 'claude mcp add backlog --scope user -- backlog mcp start' for Claude Code). A single user-scope server works across all repositories.
Which AI clients work with Backlog.md?+
Backlog.md supports Claude Code, Codex, Gemini CLI, Kiro, Cursor, and any MCP-compatible AI assistant. It also works with CLI-based agents through the backlog command-line tool and browser-based workflows through the built-in web UI.
Do I need API keys or a Git repository to use Backlog.md?+
No API keys are required—Backlog.md is fully local and includes no telemetry. Git is optional; use 'backlog init --no-git' for filesystem-only projects, though Git integration enables version-controlled task history and branch-aware features.
Is Backlog.md free and open source?+
Yes, Backlog.md is released under the MIT License, making it free for any use including commercial projects. The source code is publicly available on GitHub.
How does the MCP server find my active project?+
The Backlog.md MCP server automatically resolves the active project from your client's MCP workspace roots, following you as you switch between repositories or worktrees. You can also pin it to a specific project using the BACKLOG_CWD environment variable.
How do I install Backlog.md?+
Open the source repository on GitHub and follow its README. Backlog.md is a mcp server — MCP Agents Market links you directly to the official repo.
Is Backlog.md free?+
Backlog.md is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.