https://github.com/diazoxide/autoteam
Orchestrate AI agents with YAML-driven workflows via universal Model Context Protocol (MCP)
https://github.com/diazoxide/autoteam
agents ai aiagent chatgpt claude gemini github mcp openai orchestration
Last synced: 3 months ago
JSON representation
Orchestrate AI agents with YAML-driven workflows via universal Model Context Protocol (MCP)
- Host: GitHub
- URL: https://github.com/diazoxide/autoteam
- Owner: diazoxide
- Created: 2025-07-26T15:27:43.000Z (6 months ago)
- Default Branch: main
- Last Pushed: 2025-10-01T12:08:54.000Z (4 months ago)
- Last Synced: 2025-10-01T13:04:18.361Z (4 months ago)
- Topics: agents, ai, aiagent, chatgpt, claude, gemini, github, mcp, openai, orchestration
- Language: Go
- Homepage:
- Size: 39.3 MB
- Stars: 36
- Watchers: 0
- Forks: 4
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# AutoTeam 🤖
[](https://golang.org)
[](LICENSE)
[](https://github.com/diazoxide/autoteam/stargazers)
[](https://hub.docker.com/r/diazoxide/autoteam)
**Universal AI Agent Orchestration Platform powered by Model Context Protocol (MCP)**
[Documentation](docs/) • [Installation](docs/installation.md) • [Configuration](docs/configuration.md) • [Examples](examples/) • [Contributing](#contributing)
---
AutoTeam is a platform-agnostic orchestration system that connects AI agents with any service through MCP servers. Think of it as an **MCP hub** that enables intelligent workflows across platforms, databases, APIs, and services.
## 🎯 What is AutoTeam?
AutoTeam orchestrates AI agents (Claude Code, Gemini CLI, Qwen Code, and more) to work autonomously across any platform that supports MCP. The agent list is fully extensible - add any AI tool that fits your needs. Instead of building custom integrations, you configure MCP servers and let intelligent agents handle complex, multi-platform workflows.
### Why AutoTeam?
- 🚀 **10x Productivity**: Teams report handling 5-10x more routine tasks
- 🔗 **Universal Integration**: Connect any MCP-enabled service without custom code
- 🤝 **True Collaboration**: AI agents work in parallel, like real team members
- 📈 **Scalable Architecture**: Add agents and services as your needs grow
- 🛡️ **Enterprise Ready**: Container-native with full security isolation
## 👥 Scale Your Team with Virtual Workers
Transform your development workflow by adding AI agents as virtual team members. Each agent specializes in different roles and works in parallel, dramatically scaling your team's capacity:
```mermaid
graph TB
subgraph "Virtual Development Team"
SD[👨💻 Senior Developer
Claude Code Agent
Code reviews, Implementation]
ARCH[🏗️ Architect
Claude Code Agent
Design, Technical decisions]
QA[🧪 QA Assistant
Qwen Code Agent
Testing, Quality checks]
end
subgraph "Parallel Execution"
SD -.->|Simultaneously| FLOW1[PR Reviews
Feature Implementation]
ARCH -.->|Simultaneously| FLOW2[Architecture Review
Technical Planning]
QA -.->|Simultaneously| FLOW3[Test Automation
Quality Reports]
end
subgraph "Shared MCP Services"
GitHub_MCP[🐙 GitHub MCP
Issues, PRs, Code]
Slack_MCP[💬 Slack MCP
Communications]
DB_MCP[🗄️ Database MCP
Analytics, Metrics]
end
subgraph "Platform Integration"
GitHub[GitHub Repository]
Slack[Team Slack]
Analytics[(Analytics DB)]
end
SD --> GitHub_MCP
ARCH --> GitHub_MCP
QA --> GitHub_MCP
SD --> Slack_MCP
ARCH --> Slack_MCP
QA --> DB_MCP
GitHub_MCP --> GitHub
Slack_MCP --> Slack
DB_MCP --> Analytics
classDef agent fill:#e3f2fd,stroke:#1976d2,stroke-width:2px
classDef mcp fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px
classDef platform fill:#e8f5e8,stroke-width:2px
class SD,ARCH,QA agent
class GitHub_MCP,Slack_MCP,DB_MCP mcp
class GitHub,Slack,Analytics platform
```
**Real Impact**: Teams report handling 5-10x more routine tasks with virtual workers, allowing humans to focus on strategy and complex problem-solving.
## 📣 Marketing Team Automation
AutoTeam also scales non-technical teams. Here's how a marketing team leverages AI agents for content creation, campaign management, and analytics:
```mermaid
graph TB
subgraph "Virtual Marketing Team"
CM[📝 Content Manager
Claude Code Agent
Blog posts, Social content]
SM[📱 Social Media Manager
Gemini CLI Agent
Scheduling, Engagement]
DA[📊 Data Analyst
Qwen Code Agent
Analytics, Reports]
end
subgraph "Parallel Marketing Operations"
CM -.->|Simultaneously| MFLOW1[Content Creation
SEO Optimization]
SM -.->|Simultaneously| MFLOW2[Social Posting
Community Management]
DA -.->|Simultaneously| MFLOW3[Campaign Analysis
Performance Reports]
end
subgraph "Marketing MCP Services"
CMS_MCP[📄 CMS MCP
WordPress, Ghost]
Social_MCP[📲 Social MCP
Twitter, LinkedIn]
Analytics_MCP[📈 Analytics MCP
Google Analytics, HubSpot]
end
subgraph "Marketing Platforms"
CMS[Content Management]
SocialPlatforms[Social Networks]
MarketingTools[Analytics & CRM]
end
CM --> CMS_MCP
SM --> Social_MCP
DA --> Analytics_MCP
CM --> Social_MCP
SM --> Analytics_MCP
DA --> CMS_MCP
CMS_MCP --> CMS
Social_MCP --> SocialPlatforms
Analytics_MCP --> MarketingTools
classDef agent fill:#fff3e0,stroke:#f57c00,stroke-width:2px
classDef mcp fill:#e8f5e8,stroke:#4caf50,stroke-width:2px
classDef platform fill:#fce4ec,stroke:#e91e63,stroke-width:2px
class CM,SM,DA agent
class CMS_MCP,Social_MCP,Analytics_MCP mcp
class CMS,SocialPlatforms,MarketingTools platform
```
**Marketing Results**: Content production increased 400%, social engagement up 250%, with data-driven insights delivered daily instead of monthly.
## 🎧 Customer Support Team Automation
Scale your support operations with AI agents that handle multiple channels simultaneously, ensuring no customer request goes unnoticed:
```mermaid
graph TB
subgraph "Virtual Support Team"
SC[🎧 Support Coordinator
Claude Code Agent
Ticket triage, Escalation]
CR[💬 Chat Representative
Gemini CLI Agent
Live chat, Quick responses]
KB[📚 Knowledge Specialist
Qwen Code Agent
Documentation, Solutions]
end
subgraph "Parallel Support Operations"
SC -.->|Simultaneously| SFLOW1[Ticket Routing
Priority Assignment]
CR -.->|Simultaneously| SFLOW2[Customer Chat
Issue Resolution]
KB -.->|Simultaneously| SFLOW3[Solution Research
KB Updates]
end
subgraph "Support MCP Services"
Ticket_MCP[🎫 Ticketing MCP
Zendesk, Freshdesk]
Chat_MCP[💭 Chat MCP
Intercom, LiveChat]
KB_MCP[📖 Knowledge MCP
Confluence, Notion]
end
subgraph "Support Platforms"
HelpDesk[Help Desk System]
ChatPlatform[Live Chat Platform]
KnowledgeBase[Knowledge Base]
end
SC --> Ticket_MCP
CR --> Chat_MCP
KB --> KB_MCP
SC --> Chat_MCP
CR --> KB_MCP
KB --> Ticket_MCP
Ticket_MCP --> HelpDesk
Chat_MCP --> ChatPlatform
KB_MCP --> KnowledgeBase
classDef agent fill:#e1f5fe,stroke:#0277bd,stroke-width:2px
classDef mcp fill:#f1f8e9,stroke:#689f38,stroke-width:2px
classDef platform fill:#fef7e0,stroke:#ffa000,stroke-width:2px
class SC,CR,KB agent
class Ticket_MCP,Chat_MCP,KB_MCP mcp
class HelpDesk,ChatPlatform,KnowledgeBase platform
```
**Support Results**: 60% faster response times, 45% better escalation accuracy, 24/7 coverage with consistent service quality across all channels.
```mermaid
graph TB
subgraph "AutoTeam Core"
ATC[AutoTeam Orchestrator]
FE[Flow Engine]
WM[Worker Manager]
end
subgraph "AI Agents (Scalable)"
Claude[Claude Code Agent]
Gemini[Gemini CLI Agent]
Qwen[Qwen Code Agent]
More[...More AI Agents]
end
subgraph "MCP Servers"
GMCP[GitHub MCP]
SMCP[Slack MCP]
DMCP[Database MCP]
FMCP[Filesystem MCP]
CMCP[Custom MCP]
end
subgraph "External Platforms"
GitHub[GitHub API]
Slack[Slack API]
Database[(Database)]
FileSystem[File System]
Custom[Custom APIs]
end
ATC --> FE
ATC --> WM
FE --> Claude
FE --> Gemini
FE --> Qwen
Claude --> GMCP
Gemini --> SMCP
Qwen --> DMCP
Claude --> FMCP
Gemini --> CMCP
GMCP --> GitHub
SMCP --> Slack
DMCP --> Database
FMCP --> FileSystem
CMCP --> Custom
```
## ✨ Key Features
| Feature | Description |
|---------|-------------|
| 🌐 **Universal Platform Integration** | Connect any MCP-enabled service without custom code |
| 🔄 **Intelligent Flow Orchestration** | Parallel execution with smart dependency resolution |
| 🤖 **Multi-AI Agent Support** | Claude Code, Gemini CLI, Qwen Code, and more working together |
| 🏗️ **Container-Native Architecture** | Isolated, secure, and scalable agent deployment |
| 🚀 **Runtime Abstraction** | Flexible deployment runtime (Docker, Kubernetes, etc.) with native API control |
| ⚙️ **Configuration-Driven** | Define complex workflows in simple YAML |
| 🔌 **Extensible Plugin System** | Add custom MCP servers and AI agents |
| 📊 **Real-time Monitoring** | Track agent performance and workflow execution |
| 🎛️ **Control Plane API** | Centralized worker management with Swagger UI |
| 🔐 **Enterprise Security** | Role-based access control and secure credentials |
## 🏗️ Architecture Overview
AutoTeam acts as an intelligent MCP hub, enabling seamless communication between AI agents and platforms:
```mermaid
graph LR
subgraph "Flow Execution"
F1[Collect GitHub
Gemini CLI]
F2[Collect Slack
Claude Code]
F3[Collect Database
Qwen Code]
F4[Process All Tasks
Claude Code]
F1 --> F4
F2 --> F4
F3 --> F4
end
subgraph "MCP Connectivity"
F1 -.-> GitHub_MCP
F2 -.-> Slack_MCP
F3 -.-> DB_MCP
F4 -.-> GitHub_MCP
F4 -.-> Slack_MCP
end
GitHub_MCP --> GitHub_API[GitHub]
Slack_MCP --> Slack_API[Slack]
DB_MCP --> Database_API[(DB)]
```
## 🚀 Quick Start
### Prerequisites
- Docker 20.10+ or Podman 3.0+
- 4GB RAM minimum (8GB recommended)
- Linux, macOS, or Windows with WSL2
### 1. Install
```bash
# One-line installation
curl -fsSL https://raw.githubusercontent.com/diazoxide/autoteam/main/scripts/install.sh | bash
# Or with specific version
curl -fsSL https://raw.githubusercontent.com/diazoxide/autoteam/main/scripts/install.sh | bash -s -- --version v1.0.0
```
### 2. Initialize
```bash
# Create a new AutoTeam project
autoteam init
# Or initialize with a template
autoteam init --template development-team
```
### 3. Configure
```yaml
# autoteam.yaml
workers:
- name: "AI Assistant"
enabled: true
prompt: "Handle tasks across platforms using available MCP tools"
settings:
mcp_servers:
github:
command: /opt/autoteam/bin/github-mcp-server
args: ["stdio"]
slack:
command: /opt/autoteam/bin/slack-mcp-server
args: ["stdio"]
flow:
- name: process_tasks
type: claude
prompt: "Process tasks using MCP tools"
```
### 4. Generate Configuration (Optional)
```bash
# Generate configuration files for the runtime
autoteam generate
```
### 5. Deploy
```bash
autoteam up
```
### 6. Monitor (Optional)
```bash
# Access control plane API at http://localhost:9090
# View Swagger UI at http://localhost:9090/docs/
```
## 📚 Documentation
### Getting Started
- 📖 [Installation Guide](docs/installation.md) - Complete setup instructions
- ⚙️ [Configuration](docs/configuration.md) - Platform and agent configuration
- 🚀 [Examples](docs/examples.md) - Real-world use cases and templates
### Advanced Topics
- 🔄 [Flow System](docs/flows.md) - Workflow definition and orchestration
- 🔌 [MCP Integration](docs/mcp.md) - Platform connectivity guide
- 🏗️ [Architecture](docs/architecture.md) - System design deep dive
- 🛠️ [Development](docs/development.md) - Contributing and extending AutoTeam
### Quick Links
- [API Reference](https://pkg.go.dev/github.com/diazoxide/autoteam)
- [Control Plane API](docs/control-plane.md)
- [CLI Commands](docs/cli.md)
- [Troubleshooting](docs/troubleshooting.md)
- [FAQ](docs/faq.md)
## 💡 Use Cases
### Development Teams
- 🔍 **Code Review Automation** - Parallel PR reviews with multiple AI perspectives
- 🐛 **Issue Management** - Automatic triage, labeling, and assignment
- 🚀 **CI/CD Enhancement** - Intelligent build failure analysis and fixes
- 📝 **Documentation Generation** - Keep docs in sync with code changes
### Marketing Teams
- ✍️ **Content Production** - Blog posts, social media, email campaigns
- 📊 **Analytics Automation** - Daily reports and campaign insights
- 🎯 **SEO Optimization** - Content analysis and improvement suggestions
- 📱 **Social Media Management** - Multi-platform posting and engagement
### Customer Support
- 🎫 **Ticket Automation** - Intelligent routing and prioritization
- 💬 **Multi-Channel Support** - Unified response across chat, email, social
- 📚 **Knowledge Base Updates** - Automatic solution documentation
- 📈 **Support Analytics** - Performance metrics and trend analysis
### Data Operations
- 🔄 **ETL Pipelines** - Intelligent data transformation workflows
- 📊 **Report Generation** - Automated insights and visualizations
- 🔍 **Data Quality** - Validation and anomaly detection
- 🗄️ **Database Management** - Schema updates and optimization
## 💻 Example: Multi-Platform Workflow
```yaml
flow:
# Parallel data collection
- name: scan_github
type: gemini
prompt: "Collect urgent GitHub notifications"
- name: scan_slack
type: claude
prompt: "Check Slack for team mentions"
# Process collected data
- name: handle_tasks
type: claude
depends_on: [scan_github, scan_slack]
prompt: "Process all collected tasks with appropriate actions"
```
## 🤝 Contributing
AutoTeam is open source and welcomes contributions!
### How to Contribute
1. ⭐ **Star the repository** to show your support
2. 🐛 **Report bugs** via [GitHub Issues](https://github.com/diazoxide/autoteam/issues)
3. 💡 **Request features** in [Discussions](https://github.com/diazoxide/autoteam/discussions)
4. 🔧 **Submit pull requests** - see [Contributing Guide](CONTRIBUTING.md)
5. 📖 **Improve documentation** - even typo fixes are valuable!
6. 🔌 **Create MCP integrations** - expand the ecosystem
### Development Setup
```bash
# Clone the repository
git clone https://github.com/diazoxide/autoteam.git
cd autoteam
# Install dependencies
make deps
# Run tests
make test
# Build locally
make build
```
## 🔒 Security
For security issues, please email security@autoteam.io instead of using the issue tracker. See our [Security Policy](SECURITY.md) for more details.
## 📄 License
MIT License - see [LICENSE](LICENSE) for details.
## 🙏 Acknowledgments
- [Anthropic](https://anthropic.com) for Claude and MCP
- [Google](https://google.com) for Gemini
- [Alibaba Cloud](https://alibabacloud.com) for Qwen
- All our [contributors](https://github.com/diazoxide/autoteam/graphs/contributors)
---
**Ready to orchestrate your AI agents?**
[🚀 Get Started](docs/installation.md) • [📖 Read Docs](docs/)