Claude.mgr

Key Features

• •Multi-agent management: Launch, monitor, and manage multiple Claude Code agents simultaneously through a single spatial 3D interface. Each agent is visually represented with status indicators for idle, processing, completed, or error states, enabling quick identification of progress and issues.
• •Usage analytics: Track token consumption, API costs, and agent performance metrics per agent and overall. Dashboards and exportable reports provide clear visibility into Claude API usage, helping users optimize prompts and budget effectively without manual logging.
• •Git worktree integration: Automatically create and manage isolated Git worktrees for each agent. This keeps code changes from different agents separate, avoiding merge conflicts and supporting safe parallel development on multiple branches or versions of a codebase.
• •Real-time status monitoring: See live updates on each agent's activity, including streaming output and progress indicators. The 3D interface provides an intuitive visual overview of all running agents, reducing the need for terminal switching and log chasing.
• •Agent configuration profiles: Predefine system prompts, API keys, working directories, and other settings for different agent types. This streamlines launching agents for repeated tasks and ensures consistent behavior across sessions.
• •Built-in agent console: Interact with individual agents via a console embedded in the interface. Users can send additional instructions, inspect output in real time, or debug issues without leaving the application.

Use Cases

• •Simultaneous code refactoring: A developer launches five Claude.mgr agents, each assigned to refactor a different module of a legacy codebase. The 3D interface shows progress, and the analytics track total token usage. The outcome is parallel refactoring completed in half the time with full cost transparency.
• •Automated unit test generation: While a developer writes new feature code, Claude.mgr spawns an agent dedicated to generating unit tests. The agent works in its own Git worktree, and upon completion, provides a summary of tests created and coverage achieved, reducing manual test writing effort.
• •Multi-branch feature development: A team uses Claude.mgr to develop features on separate branches. Each branch gets an agent that codes in its own worktree. Git worktree integration ensures no branch interference, and the dashboard shows which branches are being worked on and by which agent.
• •Pull request review assistance: An open-source maintainer launches agents to analyze each incoming pull request. The agents examine code changes, run basic checks in isolated worktrees, and report findings. This speeds up review cycles and helps ensure consistency across contributions.
• •API cost optimization: A team lead uses Claude.mgr's analytics to review token consumption across all agents over a week. They identify agents with unusually high usage and adjust prompts or agent counts. This leads to a 30% reduction in monthly API costs while maintaining output quality.

Who is this for?

Claude.mgr is designed for individual developers, data scientists, and small to mid-sized teams who actively use Claude Code for coding assistance and want to scale their workflow beyond single-agent usage. It suits power users who frequently run multiple agents for parallel tasks such as code generation, refactoring, testing, and documentation. The tool is also valuable for team leads and technical managers who need visibility into AI agent utilization and costs. It requires familiarity with Claude Code and a desktop environment (macOS or Windows).