Markdown Task Manager takes a different approach to task management: it transforms plain Markdown files stored locally into an interactive Kanban board right in your browser. The standout feature isn’t just that it’s a zero-dependency, single HTML file app — it’s that it integrates deeply with AI assistants, allowing programmatic task management with full traceability.
What markdown task manager is and how it works
At its core, Markdown Task Manager is a standalone single-file HTML application. It uses the browser’s File System Access API to load and save task data in human-readable Markdown files (kanban.md for active tasks and archive.md for archived ones). This means no backend, no database, and no build step — just a modern browser (Chrome 86+, Edge 86+, Opera 72+) and local files.
The app presents your tasks as a Kanban board, where you can drag and drop cards between columns. The tasks themselves are markdown entries, making this system Git-friendly and easily editable outside the app with any text editor.
One of the most interesting design choices is that it remembers your last 10 projects and supports multi-project workflows from a single HTML file. This is handy for developers juggling multiple repos or task sets.
Under the hood, this repo relies entirely on browser technology: the File System Access API for persistent local file handling, plain HTML, CSS, and JavaScript for the UI. There’s no server or external dependency. This architecture means the app is truly local-first and privacy-respecting.
AI assistant integration: the real differentiator
What sets Markdown Task Manager apart is its AI assistant integration layer. The project ships with structured workflow files and a Claude Code skill that enables AI assistants like Claude, ChatGPT, GitHub Copilot, Gemini, and others to read and write tasks programmatically.
This works through a strict Markdown format for tasks and commands, allowing AI agents to create, update, and archive tasks with precise formatting. It means your AI assistants can document their work directly on your Kanban board, maintaining full traceability.
This approach is fairly rare and worth understanding even if you don’t immediately adopt it. Instead of treating AI helpers as black boxes, this system integrates them as first-class citizens in your project management workflow. The AI skills are shipped as markdown files defining slash commands and prompt templates for Claude Code and compatible agents.
The tradeoff here is complexity: to make use of the AI features, you need some familiarity with Claude Code or similar AI assistant tooling. Also, the system’s usefulness is limited to browsers supporting the File System Access API — Firefox and Safari are out. But the upside is a powerful, zero-setup AI-augmented Kanban board that fits naturally into a developer’s existing Markdown and Git workflows.
Quick start: getting going in three steps
The installation is minimal and requires no build or server setup:
Prerequisites
- Compatible browser: Chrome 86+, Edge 86+, or Opera 72+
- Note: File System Access API is not available on Firefox or Safari
Installation in 3 steps
- Download
task-manager.htmlfrom the repository - Open it in your browser (just double-click)
- Select a folder to store your tasks
That’s it! On first launch, the app asks you to choose a folder. If the folder is empty, it automatically creates kanban.md and archive.md files for active and archived tasks respectively.
You can name the project, and it will remember it for future sessions. This makes it very easy to start managing tasks locally without complex setup or dependencies.
Verdict: who this is for and where it fits
Markdown Task Manager is a solid pick if you want a local-first Kanban board that fits into a Markdown-based Git workflow with zero backend complexity. Its AI assistant integration layer is a compelling innovation for teams or individuals exploring AI-augmented task management, especially if you already use Claude Code or similar AI frameworks.
However, it’s not for everyone. The reliance on the File System Access API means limited browser support. Also, it’s opinionated about using Markdown files as the source of truth, which might not suit users expecting syncing or server-based task management.
The code is surprisingly clean for a single-file app, and the project thoughtfully supports multi-project workflows without user friction. The AI integration is the kind of experiment worth watching: it’s not just fluff but a concrete system allowing AI agents to manage tasks programmatically with traceability.
If you want a privacy-first, Git-friendly Kanban with a peek into AI-assisted workflows, this is worth your time.
Related Articles
- MindForger: A C++ personal knowledge manager with semantic Markdown navigation — MindForger is a cross-platform C++ desktop app combining a Markdown IDE and personal knowledge management with local sem
- How Kiln orchestrates multi-agent AI workflows using markdown and Claude Code — Kiln implements a 7-step multi-agent AI pipeline entirely through markdown files and Claude Code’s native primitives, av
- Agent Kanban: orchestrating AI coding agents with cryptographic identities — Agent Kanban is a TypeScript multi-agent platform that uses Ed25519 cryptographic identities to manage AI coding agents
- Ghist: repo-native task management for AI coding agents with conflict-free JSON storage — Ghist replaces Jira and Linear with repo-native task tracking using JSON files to avoid Git conflicts, syncing seamlessl
- abtop: process-based monitoring of AI coding agents in Rust with tmux integration — abtop is a Rust TUI that monitors local AI coding agents by scanning processes and files without API keys. Supports macO
→ GitHub Repo: ioniks/MarkdownTaskManager ⭐ 400 · HTML