Veritas Kanban is a local-first, open-source Kanban board designed for developers and AI agents. It provides a visual task management system that integrates seamlessly with autonomous coding workflows, offering a developer-centric alternative to traditional project management tools. The primary benefit is its deep integration with AI agents and Git, coupled with a database-free, Markdown-based storage approach.

How It Works

The platform utilizes a React/Vite frontend communicating with an Express backend via REST API and WebSockets. Tasks are stored as human-readable Markdown files with YAML frontmatter, eliminating the need for a traditional database. AI agents interact with Veritas Kanban through its platform-agnostic REST API to create, update, and complete tasks, with the frontend reflecting changes in real-time. This architecture enables a local-first, version-controlled workflow directly tied to AI agent execution.

Quick Start & Requirements

• Primary install/run command: git clone https://github.com/BradGroux/veritas-kanban.git && cd veritas-kanban && pnpm install && cp server/.env.example server/.env && pnpm dev
• Prerequisites: Node.js (22+), pnpm (9+).
• Estimated setup time: Under 5 minutes.
• Relevant links: Quickstart, Features, All Features, Docs, Troubleshooting, API, Agent Integration, MCP Server, Contributing, Changelog.

Highlighted Details

• Native AI agent orchestration via a platform-agnostic REST API and WebSocket interface.
• Local-first architecture with tasks stored as Markdown files, enabling direct grep-ability and version control.
• Integrated Git worktree management for isolated branches per task, with automatic cleanup.
• Bidirectional synchronization with GitHub Issues, mapping labels to task fields.
• Built-in code review, approval workflows, and visual merge conflict resolution.
• Comprehensive CLI tool (vk) for managing tasks, time tracking, agent status, and project workflows.

Maintenance & Community

The project is maintained by Brad Groux (CEO of Digital Meld). Support, bug reports, and feature requests are handled exclusively through GitHub Issues and Discussions, with a focus on community contribution.

Licensing & Compatibility

The project is licensed under the MIT License. This permissive license allows for commercial use, modification, and distribution, making it compatible with closed-source projects.

Limitations & Caveats

The project strongly emphasizes AI agent safety, warning that agents can cause significant damage if not properly guarded. Its local-first nature means users are responsible for managing data backups and synchronization if cloud access is desired. While it supports multiple agents, advanced multi-agent coordination features are listed on the roadmap.