https://github.com/dirien/agentic-ai-idp-demo
https://github.com/dirien/agentic-ai-idp-demo
Last synced: 5 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/dirien/agentic-ai-idp-demo
- Owner: dirien
- Created: 2025-05-17T18:45:18.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-09-17T18:23:50.000Z (10 months ago)
- Last Synced: 2025-10-13T08:46:37.967Z (9 months ago)
- Size: 11.7 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Agentic AI IDP Demo: Multi-Agent Infrastructure Deployment Platform
This repository demonstrates an advanced **Agentic AI Workflow** for the **Internal Developer Platform (IDP) of the Future**, powered by **Flowise AI**. The system orchestrates multiple specialized AI agents to automate cloud infrastructure deployment, documentation generation, and repository management.
## Overview
This demo showcases how AI agents can collaborate to streamline infrastructure deployment workflows, making complex cloud operations accessible through natural language interactions. The system implements "Golden AI Paths" - predefined workflows that guide users through best practices for common deployment scenarios.
## Architecture
### Flowise AI Multi-Agent System
The demo is built using **Flowise**, a low-code visual AI workflow builder that enables the creation of sophisticated multi-agent systems. Our implementation features:
- **Supervisor Agent**: Orchestrates the entire workflow and coordinates between specialized agents
- **GitHub Agent**: Handles repository operations, file management, and version control
- **Pulumi Agent**: Manages cloud infrastructure deployment and preview operations
- **README Builder Agent**: Generates comprehensive documentation
- **AWS Docs Guide Agent**: Provides contextual AWS documentation and best practices
### Agent Coordination
The agents operate using a **supervisor pattern** where:
1. The Supervisor receives user requests and delegates tasks to appropriate specialized agents
2. Each agent reports back to the Supervisor upon task completion
3. The Supervisor orchestrates the flow until all requirements are fulfilled
4. Memory persistence ensures context is maintained across the conversation
## Golden AI Paths
The system includes predefined "Golden AI Paths" that demonstrate common infrastructure deployment patterns:
### 1. Static Website on AWS S3
- **File**: `golden-ai-path-static-website-s3-aws.md`
- **Purpose**: Deploys a static website to AWS S3 with automated configuration
- **Workflow**: Repository setup → Pulumi configuration → Infrastructure deployment → Documentation generation
### 2. DigitalOcean Kubernetes Cluster
- **File**: `golden-ai-path-do-kubernetes-argocd.md`
- **Purpose**: Creates a DigitalOcean Kubernetes cluster with ArgoCD for GitOps
- **Workflow**: Blueprint retrieval → Environment setup → Infrastructure preview → Documentation
## Components
### Flowise Configuration
- **File**: `flowiseai-agentic-flow.json`
- **Contains**: Complete Flowise workflow definition with all agents, connections, and configurations
- **Models**: Utilizes AWS Bedrock with Claude 3.7 Sonnet and Claude 3.5 Sonnet models
- **Memory**: SQLite-based agent memory for conversation state management
### Infrastructure Integration
The system integrates with multiple infrastructure tools and platforms:
- **Pulumi**: Infrastructure as Code for cloud deployments
- **AWS**: Cloud provider integration with full service support
- **DigitalOcean**: Alternative cloud provider for Kubernetes deployments
- **GitHub**: Version control and repository management
- **ArgoCD**: GitOps continuous deployment
### Model Context Protocol (MCP) Servers
The agents utilize various MCP servers for enhanced capabilities:
- **GitHub MCP**: Repository operations and file management
- **Filesystem MCP**: Local file system operations
- **Git MCP**: Version control operations
- **Command MCP**: System command execution
- **AWS Documentation MCP**: Real-time AWS documentation access
- **Pulumi MCP**: Infrastructure deployment operations
## Key Features
### 🤖 **Multi-Agent Collaboration**
- Specialized agents with distinct responsibilities
- Coordinated task execution through supervisor pattern
- Persistent conversation state and context sharing
### 🛠 **Infrastructure as Code**
- Automated Pulumi template deployment
- Environment-specific configuration management
- Infrastructure preview and validation
### 📚 **Intelligent Documentation**
- Automated README generation based on deployed infrastructure
- Context-aware documentation with best practices
- Integration with live AWS documentation
### 🔄 **GitOps Integration**
- Automated repository setup and configuration
- Version-controlled infrastructure definitions
- Continuous deployment pipeline integration
### 🎯 **Golden Path Workflows**
- Predefined best-practice deployment patterns
- Step-by-step guided infrastructure setup
- Standardized organizational practices
## Getting Started
### Prerequisites
- Flowise AI instance
- AWS Account with Bedrock access
- GitHub account and personal access token
- Pulumi account and access token
- Docker (for MCP server execution)
### Setup Instructions
1. **Configure Flowise**:
- Import the `flowiseai-agentic-flow.json` configuration
- Update AWS credentials and region settings
- Configure GitHub and Pulumi access tokens
2. **Deploy a Golden Path**:
- Start a conversation with the supervisor agent
- Reference one of the Golden AI Path files
- Follow the guided workflow for your desired infrastructure
3. **Monitor Execution**:
- Track agent progress through the Flowise interface
- Review generated infrastructure configurations
- Validate deployment outputs and documentation
## Use Cases
### For Development Teams
- **Rapid Prototyping**: Quickly spin up development environments
- **Standardization**: Enforce organizational best practices through Golden Paths
- **Documentation**: Automatically generate and maintain infrastructure docs
### For Platform Engineers
- **Self-Service**: Enable developers to deploy infrastructure independently
- **Governance**: Maintain control through predefined templates and workflows
- **Observability**: Track and audit infrastructure deployment patterns
### For DevOps Teams
- **Automation**: Reduce manual infrastructure provisioning tasks
- **Consistency**: Ensure deployments follow established patterns
- **Knowledge Sharing**: Capture and distribute infrastructure expertise
## Technology Stack
- **AI Platform**: Flowise AI with Multi-Agent architecture
- **Language Models**: AWS Bedrock (Claude 3.7 Sonnet, Claude 3.5 Sonnet)
- **Infrastructure**: Pulumi, AWS, DigitalOcean
- **Version Control**: GitHub with automated repository management
- **Containerization**: Docker for MCP server execution
- **Documentation**: Automated Markdown generation
## Future Enhancements
- Integration with additional cloud providers (Azure, GCP)
- Enhanced monitoring and observability features
- Advanced security scanning and compliance checks
- Multi-environment deployment orchestration
- Custom Golden Path creation interface
## Contributing
This demo serves as a foundation for building production-ready agentic AI systems for infrastructure deployment. Contributions and extensions are welcome to expand the Golden AI Paths and agent capabilities.
---
*This project demonstrates the future of Internal Developer Platforms, where AI agents collaborate to make infrastructure deployment as simple as having a conversation.*