gastownhall / beads

A distributed graph memory layer that gives coding agents persistent, dependency-aware issue tracking across sessions and machines

维护停滞 MIT Go Tracked
28.2k 1.9k 2 小时前
CI
agents claude-code coding

星标趋势

数据积累中,暂无足够数据生成趋势图

AI 分析

项目摘要

Beads (bd) is a distributed graph issue tracker designed to give AI coding agents persistent, structured memory. Built on Dolt, it replaces ad-hoc markdown plans with a dependency-aware graph that enables agents to track, claim, and close work items across sessions and machines, preventing context loss on long-horizon tasks.

为什么值得关注

It solves a critical pain point in AI-assisted coding—agent amnesia—by providing a shared, version-controlled memory layer that multiple agents and developers can collaborate on. With 27K+ stars, 399 contributors, and integrations across Claude Code, Codex, Cursor, and other tools, it has rapidly become infrastructure for agentic coding workflows.

优势

  • Dependency-aware graph structure prevents agents from losing track of complex task relationships
  • Distributed via Dolt enables multi-machine and multi-agent collaboration with version control
  • Broad integration ecosystem with Claude Code, Codex, Cursor, Factory.ai, and other coding agents
  • Active development with 34 releases in 6 months and strong community adoption

局限性

  • 944 open issues suggests potential stability or completeness gaps despite active development
  • Relatively new project (created Oct 2025) with evolving APIs and patterns
  • Requires adoption of graph-based issue tracking paradigm which may not suit all workflows

使用场景

  • Long-horizon refactoring or feature development with AI coding agents
  • Multi-agent collaboration where different agents work on interdependent tasks
  • Persistent project memory that survives across coding sessions and machine restarts
  • Dependency management for complex codebases where task ordering matters
目标用户: Developers and teams using AI coding assistants who need agents to maintain context across long-running tasks and collaborate on shared codebases
学习曲线: 中
分析模型:LongCat-2.0 | 分析时间:24 天前