What problem does it solve?
Beads provides a persistent memory layer for AI agents, storing tasks, dependencies, and decisions in a git-backed graph. It helps agents resume multi-session work without losing context, even after conversation compaction.
Core Features & Use Cases
- Graph-based memory: Track what blocks what, relationships, and epic/subtasks.
- Compaction survivability: Context is preserved when chat history is trimmed.
- Git-backed memory: Versioned, branchable memory that teams can sync.
- Smart discovery: Quickly surface ready work with
bd ready, see blockers, and audit changes.
- Audit trails & metadata: Track status history, decisions, and rationale.
Quick Start
Install Beads CLI, initialize in your repo, and start tracking work:
- Install Beads CLI in your environment (via npm, brew, or install script).
- In your repository, run
bd init to create the .beads store.
- Create a task or epic with
bd create "Your task" -p 1 --type task.
- Use
bd ready, bd show <id>, and bd update to manage progress.