Key Features

• •Super Agent Orchestrator: An intelligent coordinator that delegates tasks among your AI agents. Give it a single objective, and it automatically assigns work to the most suitable agents, streamlining complex project execution.
• •Multi-Agent Terminal: Run several AI coding agents simultaneously with full, interactive terminal access side-by-side. This provides direct control and real-time monitoring of each agent's output and processes in one unified view.
• •Kanban Task Board: A visual drag-and-drop board for managing agent tasks. Agents can auto-pick work items based on their configured skills, creating an automated workflow from 'To Do' to 'Done' with minimal manual oversight.
• •3D Agent World: A unique animated 3D office environment where you can watch your agents 'work.' This fun, visual interface provides an engaging overview of agent activity and status at a glance.
• •Skills & Plugins: Extend agent capabilities by installing community skills from skills.sh with one click. Hundreds of pre-built skills are available to enhance what your AI agents can accomplish for specific tasks.
• •Automations: Connect to external tools like GitHub, JIRA, Telegram, and Slack. Agents can automatically process issues, messages, and tasks from these platforms, integrating AI directly into your existing workflows.
• •Multi-Provider Support: Works with popular AI models including Claude (Anthropic), Codex (OpenAI), Gemini (Google), and Ollama for local LLMs. Mix and match different agents per project for optimal results.
• •Usage Analytics: Gain insights into how your AI agents are being utilized. Monitor performance, track task completion, and analyze efficiency to optimize your multi-agent setup and resource allocation.

Use Cases

• •Development Team Coordination: A software team uses Dorothy to manage multiple AI agents assigned to different microservices. The Super Agent orchestrates code reviews and bug fixes, while the Kanban board visualizes progress across the entire project.
• •Automated Issue Triage: Connect Dorothy to a GitHub repository. When new issues are filed, agents automatically pick them up based on skill tags (e.g., 'frontend bug', 'backend API'), beginning investigation and even suggesting fixes without manual assignment.
• •Multi-Model Coding Projects: A developer working on a complex feature uses Claude for architecture design and Codex for boilerplate code generation simultaneously via the Multi-Agent Terminal, comparing outputs and merging the best results in real time.
• •CI/CD Pipeline Integration: Incorporate AI agents into a continuous integration workflow. Agents can automatically run tests from a connected JIRA ticket, report results back to Slack, and even commit passing code changes, reducing manual DevOps overhead.
• •Community Skill Deployment: A data scientist installs a specialized 'data cleaning' skill from the community library. They then assign a batch of messy datasets to an agent, which processes them automatically using the new plugin, saving hours of manual scripting.
• •Local LLM Experimentation: A researcher uses Dorothy with Ollama to run and compare several fine-tuned local language models on a coding benchmark. The 3D Agent world visually shows each model's activity and completion status during the experiment.

Who is this for?

Dorothy is built for developers, engineering teams, and tech-savvy individuals who regularly use AI coding assistants like Claude, ChatGPT (Codex), Gemini, or local models (via Ollama) in their workflow. It targets users who find themselves managing multiple AI agents for different tasks and are seeking a centralized, visual, and automated way to control them. The tool is ideal for those who want to integrate AI capabilities with existing development tools like GitHub and JIRA, or communication platforms like Slack and Telegram, to create seamless automated workflows. As a free and open-source application, it also appeals to hobbyists, researchers, and anyone interested in experimenting with multi-agent AI systems without financial or account barriers.