Overview

Beads (bd) is a revolutionary distributed graph issue tracker designed to empower your AI coding agents. It provides a robust, version-controlled memory that moves beyond simple markdown lists, offering a dependency-aware graph structure. This allows AI agents to tackle complex, long-horizon tasks with unparalleled context and continuity, preventing critical information loss.

Key Features

• Dolt-Powered Persistence: Leverages Dolt, a SQL database with Git-like versioning, for native branching, cell-level merging, and built-in synchronization.
• Agent-Optimized Design: Emits JSON output, automatically detects dependency-free tasks, and tracks task readiness for seamless AI integration.
• Zero Conflict IDs: Utilizes hash-based IDs (bd-a1b2) to eliminate merge conflicts in collaborative or multi-branch agent workflows.
• Intelligent Compaction: Implements semantic memory decay by summarizing old, closed tasks, optimizing context window usage.
• Rich Messaging & Threading: Supports issue types with built-in threading, ephemeral lifecycles, and message delegation for advanced communication.
• Knowledge Graph Linking: Enables sophisticated task relationships with relates_to, duplicates, supersedes, and replies_to link types.
• Flexible Storage: Offers both embedded (default, single-writer) and server modes (multi-writer) for diverse deployment needs.
• Git-Free Operation: Fully functional without Git, making it ideal for non-Git VCS, monorepos, CI/CD pipelines, and sandboxed environments.

Typical Use Cases

Beads is engineered for scenarios demanding persistent, structured memory for AI agents:

• Complex Codebase Development: Enables AI agents to manage intricate feature development, bug fixes, and refactoring across large projects by maintaining a clear, versioned task history.
• Long-Term AI Project Management: Facilitates AI-driven research and development by providing a stable, evolving memory that remembers progress and dependencies over extended periods.
• Automated Code Review & Analysis: Allows AI agents to track findings, suggestions, and their resolution status in a structured, traceable manner.
• Multi-Agent Collaboration: Provides a shared, version-controlled "brain" for multiple AI agents working on the same project, ensuring synchronized understanding and progress.
• Personal Coding Assistants: Enhances individual developer productivity by offering a powerful, context-aware task management system that integrates seamlessly with AI.