Dorothy  by Charlie85270

AI agent orchestrator desktop app

Created 2 months ago
256 stars

Top 98.5% on SourcePulse

GitHubView on GitHub
Project Summary

Dorothy is a desktop application designed to orchestrate multiple AI agents (including Claude Code, Codex, Gemini, and local models) from a unified interface. It targets engineers, researchers, and power users who need to manage and automate complex AI workflows, overcoming the limitations of single-instance CLI agent tools by enabling parallel execution, visual task management, and remote control.

How It Works

Dorothy employs an Electron framework with a React/Next.js frontend and a Node.js backend. It leverages node-pty to run multiple AI agents concurrently in isolated PTY terminal sessions. The core architecture relies on a suite of MCP (Model Context Protocol) servers, including an orchestrator, Telegram/Slack bots, Kanban manager, Vault, and SocialData connector, which communicate via standard input/output. This design facilitates seamless agent interaction, automation triggering, and data persistence.

Quick Start & Requirements

  • Installation: Download the latest release from GitHub Releases or build from source.
  • Prerequisites: Node.js 18+, npm/yarn, Claude Code CLI (npm install -g @anthropic-ai/claude-code), GitHub CLI (gh). Google Workspace integration requires gcloud CLI and @googleworkspace/cli.
  • Links: GitHub Releases, Claude Code CLI, Google Workspace CLI. Development server runs at http://localhost:3000.

Highlighted Details

  • Parallel Agent Management: Run an unlimited number of concurrent agents, each in its own isolated PTY terminal, supporting multi-repo contexts and Git worktree.
  • Super Agent (Orchestrator): A meta-agent that programmatically delegates, monitors, and coordinates tasks across the agent pool.
  • Automations: Trigger AI agents autonomously based on external events from sources like GitHub and JIRA, enabling CI/CD-like workflows.
  • Vault: A persistent document storage system with full-text search (SQLite FTS5) for shared knowledge bases accessible by any agent.
  • Remote Control: Manage agents and tasks via Telegram and Slack integrations.
  • Google Workspace Integration: Direct access to Gmail, Drive, Sheets, Docs, and Calendar via the gws CLI.

Maintenance & Community

Contributions are welcomed via Pull Requests. Specific details on active maintainers, sponsorships, or community channels (like Discord/Slack) are not detailed in the provided documentation.

Licensing & Compatibility

The project is licensed under the MIT License, which permits broad use, including commercial applications and integration into closed-source projects.

Limitations & Caveats

While the core functionality is robust, some integrations like Pipedrive, RSS, and custom webhooks for Automations are listed as "Planned." The README notes a specific macOS issue where downloaded applications may be flagged as "damaged," requiring a xattr -cr command to resolve. The project relies on several external CLIs, which must be installed and configured separately.

Health Check
Last Commit

1 week ago

Responsiveness

Inactive

Pull Requests (30d)
14
Issues (30d)
8
Star History
169 stars in the last 30 days

Explore Similar Projects

Starred by Andrew Ng Andrew Ng(Founder of DeepLearning.AI; Cofounder of Coursera; Professor at Stanford), Jack Lukic Jack Lukic(Author of Semantic UI), and
5 more.

ag2 by ag2ai

0.7%
4k
AgentOS for building AI agents and facilitating multi-agent cooperation
Created 1 year ago
Updated 13 hours ago
Feedback? Help us improve.