Skip to content

SETUP.md

Latest commit

 

History

History
220 lines (173 loc) · 5.3 KB

SETUP.md

File metadata and controls

220 lines (173 loc) · 5.3 KB

🚀 FlowTech-AI Setup Guide

Prerequisites

System Requirements

  • OS: Linux (Ubuntu 20.04+ recommended)
  • RAM: Minimum 8GB, Recommended 16GB+
  • Storage: Minimum 20GB free space
  • CPU: x86_64 architecture
  • Network: Internet access for initial setup

Required Software

  • Docker: Version 24.0+ with Docker Compose plugin
  • Git: For cloning the repository
  • curl: For health checks

External Dependencies

  • Ollama: Local LLM engine (install separately)

Quick Start

1. Clone Repository

git clone https://github.com/FlowTech-Lab/FlowTech-AI.git
cd FlowTech-AI

2. Run Initialization

chmod +x init.sh
sudo ./init.sh

The init.sh script will:

  • ✅ Verify system prerequisites
  • ✅ Check available disk space (minimum 2GB)
  • ✅ Download all Docker images
  • ✅ Create data directories with proper permissions
  • ✅ Configure environment variables
  • ✅ Start all services in correct order
  • ✅ Verify service health

3. Verify Installation

# Check all services are running
docker compose ps

# Test service endpoints
curl -s http://localhost:8081  # OpenWebUI
curl -s http://localhost:5678  # n8n
curl -s http://localhost:8082  # SearxNG
curl -s http://localhost:3300  # Langfuse
curl -s http://localhost:6333  # Qdrant

Detailed Setup OPTIONNAL

Ollama Installation

# Install Ollama
curl -fsSL https://ollama.ai/install.sh | sh

# Start Ollama service
ollama serve

# Pull recommended models (in another terminal)
ollama pull qwen3:4b #chat modèle for low vram memory
ollama pull qwen3:8b #chat modèle for 6gb+ Vram
ollama pull bge-m3:567m #chat embrinding modèle.

Manual Configuration

1. Environment Variables

The init.sh script creates a .env file with default values. Key variables:

# Database
POSTGRES_USER=n8n
POSTGRES_PASSWORD=<auto-generated>
POSTGRES_DB=n8n

# Services
OPENWEBUI_PORT=8081
N8N_PORT=5678
SEARXNG_PORT=8082
LANGFUSE_PORT=3300

# Langfuse
LANGFUSE_INIT_USER_EMAIL=admin@flowtech.local
LANGFUSE_INIT_USER_PASSWORD=<auto-generated>

2. Service Configuration

OpenWebUI Setup:

  1. Access: http://localhost:8081
  2. Go to Admin Panel > Settings > Web Search
  3. Configure: http://searxng:8080/search

Langfuse Setup:

  1. Access: http://localhost:3300
  2. Login with credentials from .env
  3. Create organization and project
  4. Generate API keys in Project > Settings > API Keys

n8n Workflow Setup:

  1. Access: http://localhost:5678
  2. Import main workflow: SRC/N8N-openwebui-workflow.json
  3. Import function: SRC/function-N8N Pipe.json
  4. Configure webhook URL in OpenWebUI: http://n8n:5678/webhook/invoke_n8n_agent

Network Configuration

Internal Docker Communication

Services communicate using container names:

  • PostgreSQL: postgres:5432
  • Redis: redis:6379
  • Qdrant: http://qdrant:6333
  • MinIO: minio:9000
  • ClickHouse: clickhouse:8123

External Access

For LAN access, replace localhost with your server IP:

  • OpenWebUI: http://YOUR_IP:8081
  • n8n: http://YOUR_IP:5678
  • SearxNG: http://YOUR_IP:8082
  • Langfuse: http://YOUR_IP:3300

Troubleshooting

Common Issues

PostgreSQL Volume Mount Error

If you encounter change mount propagation through procfd error:

  • The setup now uses Docker named volumes by default
  • This issue should be resolved in the current version

Permission Issues

# Fix AI_Data permissions
sudo chown -R $USER:$USER AI_Data/
chmod -R 755 AI_Data/

Service Won't Start

# Check service logs
docker compose logs [service-name]

# Restart specific service
docker compose restart [service-name]

# Complete reset (WARNING: deletes all data)
# Edit init.sh: set DEV_MODE=true
sudo ./init.sh

Port Conflicts

If default ports are in use, modify .env:

OPENWEBUI_PORT=8082
N8N_PORT=5679
# ... etc

Health Checks

# All services status
docker compose ps

# Individual service health
docker compose exec postgres pg_isready
docker compose exec redis redis-cli ping
curl -f http://localhost:6333/health  # Qdrant

Production Deployment

Security Considerations

  1. Change default passwords in .env
  2. Enable HTTPS with reverse proxy (nginx/traefik)
  3. Configure firewall to restrict port access
  4. Set up backups for AI_Data/ directory
  5. Enable monitoring with Prometheus/Grafana

Backup Strategy

# Backup data directory
tar -czf flowtech-ai-backup-$(date +%Y%m%d).tar.gz AI_Data/

# Backup PostgreSQL
docker compose exec postgres pg_dump -U n8n n8n > backup-$(date +%Y%m%d).sql

# Backup Qdrant
docker compose exec qdrant qdrant-cli snapshot create

Scaling Considerations

  • Vertical scaling: Increase VM resources (RAM/CPU)
  • Horizontal scaling: Consider Kubernetes for multiple nodes
  • Storage scaling: Use external storage for AI_Data/

Support

Logs Location

  • Application logs: logs/init-YYYYMMDD-HHMMSS.log
  • Docker logs: docker compose logs [service]
  • Service logs: Inside containers at /var/log/

Getting Help

  1. Check logs first: docker compose logs [service]
  2. Verify prerequisites are met
  3. Review this documentation
  4. Check GitHub issues
  5. Create new issue with logs and system info

Next Steps: After successful setup, see USER_GUIDE.md for usage instructions.