TFGrid AI Stack

AI-powered development platform with integrated Git hosting and deployment.

Overview

TFGrid AI Stack is an app that deploys a complete AI-powered development platform on a single VM. It automatically installs and configures three integrated components:

• Nginx Reverse Proxy: Public API gateway with routing, SSL termination, and monitoring
• AI Agent Service: AI-powered project creation and code generation (Node.js API)
• Gitea Git Server: Git repository hosting with web interface

All components run on the same VM and are accessible through clean URLs on a single domain.

Quick Start

Deploy the complete AI development environment:

tfgrid-compose up tfgrid-ai-stack

Features

• 🤖 AI-Powered Development: Create projects with AI assistance
• 📦 Integrated Git Hosting: Built-in Gitea for version control
• 🌐 Public Deployment: Automatic web deployment with SSL
• 📊 Monitoring: Prometheus + Grafana dashboards
• 🔒 Secure Networking: WireGuard VPN for inter-VM communication
• ⚡ One-Click Deployment: Single command setup
• 🔄 Automated Backups: Scheduled backups with retention policies

Architecture

Internet
    ↓
[Single VM]
├── Nginx Reverse Proxy ← routing, SSL, monitoring
├── AI Agent Service ← project creation & code generation
└── Gitea Git Server ← repository hosting & web interface

URL Structure

https://yourdomain.com/
├── /                    → Gateway dashboard/API
├── /git/               → Gitea web interface
├── /api/               → AI Agent API endpoints
├── /monitoring/        → Prometheus/Grafana
└── /project1/          → AI-generated projects

Usage

Deploy

tfgrid-compose up tfgrid-ai-stack

Create a Project

tfgrid-compose exec tfgrid-ai-stack create "portfolio website"

List Projects

tfgrid-compose exec tfgrid-ai-stack projects

Access Services

# Get deployment URLs
tfgrid-compose address tfgrid-ai-stack

# SSH into VMs
tfgrid-compose ssh tfgrid-ai-stack

Management Commands

# Monitor project logs
tfgrid-compose exec tfgrid-ai-stack monitor <project-name>

# Delete a project
tfgrid-compose exec tfgrid-ai-stack delete <project-name>

# Manual backup
tfgrid-compose exec tfgrid-ai-stack backup

# Restore from backup
tfgrid-compose exec tfgrid-ai-stack restore <backup-file>

Configuration

The pattern supports extensive customization through variables:

Domain & SSL

• domain: Custom domain name for public access
• ssl_email: Email for SSL certificate (required if domain set)

Resource Allocation

• vm_cpu/memory/disk: Single VM resources (all services co-located)

Security & Limits

• api_rate_limit: API rate limiting
• max_concurrent_projects: Concurrent project creation limit

Backup Settings

• backup_retention_days: Backup retention period
• backup_schedule: Cron schedule for automated backups

Requirements

• ThreeFold Grid account with sufficient TFT
• tfgrid-compose CLI installed
• SSH key configured

Resources

• CPU: 8 cores total (default)
• Memory: 16GB total (default)
• Disk: 200GB total (default)
• Cost: ~$15-20/month (varies by node pricing)

App Structure

tfgrid-ai-stack/
├── tfgrid-compose.yaml    # App definition (single-vm pattern)
├── deployment/            # Component orchestration hooks
│   ├── setup.sh          # Install all components on single VM
│   ├── configure.sh      # Configure nginx routing
│   └── healthcheck.sh    # Verify component health
├── scripts/              # AI project management scripts
├── README.md             # This file
└── LICENSE               # Apache 2.0

Development

This app is maintained by TFGrid Studio. It integrates components from:

• tfgrid-ai-agent - AI coding assistant logic
• tfgrid-gitea - Git hosting setup
• Single-VM pattern from tfgrid-compose - Infrastructure

Documentation

• Pattern Documentation
• API Reference
• Troubleshooting

Support

This is an official TFGrid Studio application. For support:

• Documentation: https://docs.tfgrid.studio
• Issues: https://github.com/tfgrid-studio/tfgrid-compose/issues