beads: what it is, what problem it solves & why it's gaining traction

What it solves

Beads provides a structured, persistent memory and issue tracking system specifically designed for AI coding agents. It replaces unstructured markdown files (like TODO lists) with a dependency-aware graph, preventing agents from losing context during long-horizon tasks.

How it works

Beads uses Dolt as its backend, which is a version-controlled SQL database. This allows the project to treat issues and memories as data that can be branched, merged, and synced across machines. It provides a CLI tool (bd) that agents can use to create, claim, and close tasks, and to store "insights" via a memory system. It integrates with various AI agents (such as Claude Code, Codex, and Factory.ai) by providing specific setup hooks and guidance files (AGENTS.md).

Who it’s for

Software developers and AI agent users who want to give their coding agents a reliable way to track tasks, manage dependencies between tasks, and manage project-specific knowledge without cluttering the main codebase with temporary markdown files.

Highlights

• Dolt-Powered: Version-controlled SQL database with native branching and cell-level merging.
• Agent-Optimized: Provides JSON output and auto-ready task detection for AI consumption.
• Dependency Tracking: Supports graph links like relates_to, duplicates, and supersedes to build a knowledge graph of tasks.
• Persistent Memory: The bd remember command allows agents to store insights that are later injected back into the agent's context via bd prime.
• Zero Conflict: Uses hash-based IDs to prevent merge collisions in multi-agent environments.
• Flexible Storage: Supports both embedded mode (local file) and server mode (external SQL server) for concurrent writers.