🏠 AI LaunchKit - Local Network Edition

Your Complete AI Stack for Local Networks

40+ AI services accessible via IP:PORT - Perfect for learning, development, testing, and private team collaboration without domain/SSL complexity

Quick Start • Service Ports • Network Configuration • Troubleshooting

---

🎯 What is AI LaunchKit Local?

This is a port-based local network version of the AI LaunchKit that runs completely in Docker containers without requiring domain configuration, SSL certificates, or host system modifications.

Perfect for:

• 🎓 Learning & Experimentation - Safe environment to learn AI technologies
• 💼 Professional Use - Private AI stack for teams and organizations
• 🧪 Development & Testing - Full-featured environment for building AI applications
• 🏢 Local Deployment - Keep all data on-premises in your network
• 👥 Team Collaboration - Share AI services with colleagues on same network

Key Differences from Original:

• ✅ No Domains Required - Access via IP:PORT (e.g., 192.168.1.100:8000)
• ✅ No SSL Setup - HTTP-only for local network use
• ✅ No Caddy/Reverse Proxy - Direct port access to services
• ✅ No Host Modifications - Everything runs in Docker containers
• ✅ Local Network Ready - Designed for LAN access from multiple devices
• ✅ Production-Grade - Same services as cloud version, just local

🚀 Quick Start

Prerequisites

1. Server: Ubuntu 24.04 LTS (64-bit) with fresh installation
2. Resources:
   • Minimum: 4 GB RAM, 2 CPU cores, 30GB disk
   • Recommended: 16+ GB RAM, 8+ CPU cores, 120GB disk
3. Network: Local network access (no internet domain needed)
4. Docker Management: Portainer will be installed automatically if not present

One-Command Installation

# Clone and install in one command
git clone https://github.com/heinrichhermann/ai-launchkit-local && cd ai-launchkit-local && sudo bash ./scripts/install_local.sh

Installation time: 10-15 minutes (plus optional n8n workflows import)

What Gets Installed

The installer will:

1. ✅ Update system and install Docker
2. ✅ Generate secure passwords and API keys
3. ✅ Configure services for local network access
4. ✅ Start selected services with port mappings
5. ✅ Generate access report with all URLs

No domain prompts, no SSL setup, no host system changes!

---

📖 Step-by-Step Installation Guide (For Beginners)

Perfect for: People new to Linux, AI, or Docker Time needed: 30-45 minutes Experience level: No prior knowledge required!

What You Need:

✅ A Server or Computer running Ubuntu

• Minimum: 4 GB RAM, 2 CPU cores, 30GB disk
• Recommended: 8+ GB RAM, 4+ CPU cores
• Fresh Ubuntu 24.04 LTS installation

✅ Access to the Server

• SSH connection OR direct console access
• Administrator (sudo) rights

✅ Internet Connection

• For downloading Docker images and services

❌ What You DON'T Need:

• Domain name or website
• SSL certificates
• Advanced Linux knowledge
• Programming experience

---

Step 1: Connect to Your Ubuntu Server

What happens: You'll open a terminal and connect to your server

📖 Detailed Connection Instructions

Option A: Using SSH (Remote Connection)

If your server is in another room or location:

1. Find your server's IP address

   • Usually printed during Ubuntu installation
   • Or check your router's connected devices
   • Example: 192.168.1.100

2. Open Terminal/SSH Client:

   • Windows: Download PuTTY or use Windows Terminal
   • Mac: Open Terminal app (⌘+Space, type "Terminal")
   • Linux: Open Terminal (Ctrl+Alt+T)

3. Connect via SSH:

   ssh your-username@192.168.1.100

   Replace:

   • your-username with your Ubuntu username
   • 192.168.1.100 with your server's IP

4. Enter your password when prompted

   • You won't see characters while typing - this is normal!
   • Press Enter when done

5. You're connected! You should see:

   username@server:~$
   

Option B: Direct Console Access

If you're sitting at the server:

1. Login with your username and password
2. You'll see the command prompt:

   username@server:~$
   

3. You're ready!

---

Step 2: Download and Start Installation

What happens: One command downloads everything and starts the automated installation

Copy and paste this into your terminal:

git clone https://github.com/heinrichhermann/ai-launchkit-local && cd ai-launchkit-local && sudo bash ./scripts/install_local.sh

Press Enter, then type your password when asked.

📖 What This Command Does

Breaking down the command:

Part 1: git clone https://github.com/heinrichhermann/ai-launchkit-local

• Downloads the AI Learning Kit to your server
• Creates folder: ai-launchkit-local

Part 2: cd ai-launchkit-local

• Changes into that folder
• All scripts are now accessible

Part 3: sudo bash ./scripts/install_local.sh

• sudo = Run as administrator (needed for installation)
• bash = Execute the script
• ./scripts/install_local.sh = The installation script

Why sudo?

• Installs system packages (Docker, etc.)
• Configures firewall
• Needs administrator privileges

---

Step 3: Answer Installation Questions

What happens: The wizard asks you a few simple questions to customize your installation

📖 Question-by-Question Guide

Question 1: Configure LAN Access?

Allow access from local network (recommended for LAN installation)?
Configure LAN access now? (Y/n):

What to do: ✅ Press ENTER (=Yes)

What it means:

• Opens firewall ports for your local network
• Devices on your network can access the services
• Example: Access from your laptop, phone, or tablet

Technical details: Adds firewall rules for 192.168.x.x and 10.x.x.x networks

---

Question 2: Service Selection

Choose services for your local network AI development stack.
[Checkbox menu appears]

What to do:

1. Use ↑↓ Arrow keys to move up/down
2. Press SPACEBAR to select/deselect services
3. Press ENTER when done

Recommended for beginners:

• ✅ n8n (Workflow Automation) - Already selected
• ✅ flowise (AI Agent Builder) - Already selected
• ✅ monitoring (Grafana dashboards) - Already selected
• ⚠️ Leave others unselected for now (you can add later)

What each service does:

• n8n: Build AI automation workflows (like IFTTT for AI)
• Flowise: Create chatbots with drag-and-drop
• monitoring: See system performance (RAM, CPU usage)

---

Question 3: LAN IP Address

✅ Detected LAN IP address: 192.168.1.100
Use this IP address for network access? (Y/n):

What to do: ✅ Press ENTER (=Yes)

What it means:

• Your services will be accessible at http://192.168.1.100:PORT
• This IP was automatically detected
• You can change it later if needed

---

Question 4: Ollama Hardware Profile

Choose the hardware profile for Ollama local LLM runtime.
( ) CPU (Recommended for most users)
( ) NVIDIA GPU (Requires NVIDIA drivers)
( ) AMD GPU (Requires ROCm drivers)

What to do:

• ✅ Select "CPU" and press ENTER (for most users)
• Only choose GPU options if you have a powerful graphics card

What it means:

• CPU: Uses processor for AI (slower but works everywhere)
• GPU: Uses graphics card (much faster but needs setup)

---

Question 5: Import n8n Workflows?

Import workflows? (y/n):

What to do: ⚠️ Type n and press ENTER (skip for now)

Why skip:

• Takes 30 minutes to import 300+ workflows
• You probably want to explore first
• You can import later anytime

If you choose yes:

• Installation continues in background
• Check progress: docker logs n8n-import

---

Question 6: Admin Credentials (If Monitoring or Langfuse Selected)

Admin Email: 
Admin Password:

What to do: ✅ Enter your email and a strong password

What are these for:

• Grafana Login - View system monitoring dashboards (Port 8003)
• Langfuse Login - Track LLM performance and usage (Port 8096)
• Same credentials work for both services
• This is YOUR admin account, not public-facing

Security note:

• Only accessible from your local network
• No internet exposure
• Strong password still recommended

---

Question 7: n8n Worker Count

Enter the number of n8n workers to run (default is 1):

What to do: ✅ Press ENTER (use default: 1)

What it means:

• Workers process workflows in parallel
• 1 worker is fine for learning
• More workers = can run more workflows simultaneously

---

Question 8-10: API Keys (Optional)

OpenAI API Key (press Enter to skip):
Anthropic API Key (press Enter to skip):
Groq API Key (press Enter to skip):

What to do: ✅ Press ENTER 3 times (skip all)

What are these:

• Optional paid API keys for cloud AI services
• Not needed for local AI with Ollama
• You can add them later in the .env file

---

Installation Progress:

You'll now see:

========== STEP 1: System Preparation ==========
✅ System preparation complete!

========== STEP 2: Installing Docker ==========
✅ Docker installation complete!

========== STEP 3: Generating Local Network Configuration ==========
✅ Local network configuration complete!

[... more steps ...]

🎉 AI LaunchKit Local Network Installation Complete!

Wait for all steps to complete (10-15 minutes)

---

Step 4: Services Are Ready! 🎉

What happens: All services are running, you can start using them immediately!

You'll see a summary like this:

🎉 INSTALLATION SUCCESSFUL!

Your services are accessible at:
- n8n (Workflow Automation): http://192.168.1.100:8000
- Flowise (AI Agent Builder): http://192.168.1.100:8022
- Grafana (Monitoring): http://192.168.1.100:8003
- Mailpit (Email Testing): http://192.168.1.100:8071

What to do next:

1. Open your browser (can be on any device!)
2. Navigate to n8n: http://YOUR-SERVER-IP:8000
3. Create your admin account:
   • Enter your email
   • Choose a strong password
   • Click "Next"
4. You're in! Welcome to n8n 🎉

---

🎯 Your First AI Automation (10 Minutes Tutorial)

Let's build a simple AI workflow to test everything:

📖 Complete Beginner Tutorial

Tutorial: Ask Ollama a Question

What you'll learn:

• How to create a workflow in n8n
• How to connect to Ollama (local AI)
• How to see AI responses

Steps:

1. Open n8n: http://YOUR-SERVER-IP:8000

2. Click the big "+" button (Create New Workflow)

3. Click "Add first step"

   • A panel opens on the right

4. Search for "Manual Trigger"

   • Type "manual" in the search box
   • Click "Manual Trigger" when it appears
   • This lets you start the workflow with a button click

5. Click the "+" icon next to your Manual Trigger node

   • This adds the next step

6. Search for "HTTP Request"

   • Type "http" in the search box
   • Click "HTTP Request"

7. Configure the HTTP Request: Click on the HTTP Request node and fill in:

   • Method: Select "POST" from dropdown

   • URL: Enter this exactly:

     http://ollama:11434/api/generate
     

   • Authentication: Leave as "None"

   • Send Body: Toggle ON

   • Body Content Type: Select "JSON"

   • Specify Body: Select "Using JSON"

   • JSON: Paste this:

     {
       "model": "qwen2.5:7b-instruct-q4_K_M",
       "prompt": "Hello! Please introduce yourself and explain what you can do.",
       "stream": false
     }

8. Click "Execute Workflow" button (top right)

   • Wait 5-10 seconds for AI to respond

9. See the result!

   • Click on the HTTP Request node
   • Look at the "Output" tab
   • You'll see the AI's response in JSON format

Congratulations! 🎉 You just created your first AI workflow!

What just happened:

• n8n sent a request to Ollama (your local AI)
• Ollama processed your question
• Ollama sent back an answer
• You can now build on this to create complex automations!

Next steps:

• Try different prompts
• Add more nodes (email, database, etc.)
• Explore the 300+ workflow templates

---

❓ Frequently Asked Questions

🖥️ Can I install this on Windows or Mac directly?

No, not directly. The AI Kit requires Ubuntu Linux.

Your options:

Option 1: Virtual Machine (Easiest for Windows/Mac)

1. Install VirtualBox (free)
2. Download Ubuntu 24.04 LTS
3. Create VM with 8GB RAM, 4 CPU cores, 60GB disk
4. Install Ubuntu in the VM
5. Follow installation guide above
6. Access services from your Windows/Mac browser

Option 2: Cloud VPS (No local hardware needed)

1. Rent Ubuntu server from providers like:
   • DigitalOcean ($12/month for 4GB RAM)
   • Linode/Akamai ($10/month for 4GB RAM)
   • Hetzner ($5/month for 4GB RAM - Europe)
2. Get SSH access details
3. Follow installation guide above
4. Access from anywhere

Option 3: Raspberry Pi (Cheap dedicated hardware)

• Raspberry Pi 4/5 with 8GB RAM ($80)
• Install Ubuntu Server 24.04
• Perfect for 24/7 operation
• Low power consumption

❌ I got an error during installation, what should I do?

Don't worry! The system automatically cleans up on errors.

Steps to resolve:

1. Read the error message carefully

   • It shows exactly which step failed
   • Example: "Installation failed at: Docker Installation"

2. Check common issues:

   Error: "Port already in use"

   • Another program is using that port
   • Check what: sudo netstat -tuln | grep :8000
   • Stop that program or choose different port

   Error: "Out of memory" or "Cannot allocate memory"

   • Server doesn't have enough RAM
   • Need minimum 4GB RAM
   • Upgrade server or select fewer services

   Error: "Docker daemon not responding"

   • Docker installation failed
   • Try: sudo systemctl status docker
   • Restart: sudo systemctl restart docker

   Error: "Permission denied"

   • Not running with sudo
   • Use: sudo bash ./scripts/install_local.sh

3. Try installation again:

   cd ai-launchkit-local
   sudo bash ./scripts/install_local.sh

   • Previous failed installation was rolled back automatically
   • Safe to re-run

4. Still having issues?

   • Post in GitHub Issues
   • Include: Error message, Ubuntu version, server specs
   • Community will help!

📱 How do I access services from my phone, laptop, or tablet?

It's automatic! No configuration needed.

Steps:

1. Make sure your device is on the same WiFi/network as the server

2. Find your server's IP (shown during installation)

   • Example: 192.168.1.100

3. Open any browser on your device

   • Chrome, Safari, Firefox, Edge - all work!

4. Type in the address bar:

   http://192.168.1.100:8000
   

   Replace 192.168.1.100 with YOUR server IP

5. Done! Services work identically on all devices

Examples:

• From laptop: http://192.168.1.100:8000 (n8n)
• From phone: http://192.168.1.100:8022 (Flowise)
• From tablet: http://192.168.1.100:8003 (Grafana)
• From any device: http://192.168.1.100/ (Dashboard)

Tip: Bookmark these URLs on your devices!

🗑️ How do I uninstall everything?

Safe and simple uninstallation with automatic backup:

cd ai-launchkit-local
sudo bash ./scripts/uninstall_local.sh

What happens:

1. Shows current status

   • Lists all running services
   • Shows data volumes

2. Asks for confirmation

   • Type yes to proceed
   • Anything else cancels

3. Offers backup (Recommended!)

   • Creates backup of workflows, databases, volumes
   • Saves to: ~/ai-launchkit-backup-YYYYMMDD-HHMMSS/
   • Press ENTER to accept

4. Removes AI LaunchKit

   • Stops all containers
   • Removes data (with your permission)
   • Cleans up Docker images

5. Preserves important stuff:

   • ✅ Docker (in case you use it for other things)
   • ✅ Portainer (Docker management tool)
   • ✅ Your project folder (can reinstall anytime)

After uninstall:

• Server is back to clean state
• Can reinstall anytime: sudo bash ./scripts/install_local.sh
• Or restore from backup

🔧 What if I want to change the services later?

You can easily add or remove services:

1. Stop all services:

   cd ai-launchkit-local
   docker compose -p localai -f docker-compose.local.yml down

2. Edit the .env file:

   nano .env

3. Find the line starting with: COMPOSE_PROFILES=

4. Add or remove service names (comma-separated)

   • Example: COMPOSE_PROFILES="n8n,flowise,monitoring,cpu,comfyui"
   • Available: n8n, flowise, bolt, openui, monitoring, cpu, gpu-nvidia, gpu-amd, calcom, baserow, nocodb, vikunja, leantime, and more

5. Save and exit:

   • Press Ctrl+X
   • Press Y (yes, save)
   • Press Enter

6. Start services again:

   docker compose -p localai -f docker-compose.local.yml up -d

7. Wait 2-3 minutes for new services to start

⚡ How do I stop or restart services?

Stop all services:

cd ai-launchkit-local
docker compose -p localai -f docker-compose.local.yml down

Start all services:

cd ai-launchkit-local
docker compose -p localai -f docker-compose.local.yml up -d

Restart a specific service (example: n8n):

docker compose -p localai -f docker-compose.local.yml restart n8n

See what's running:

docker ps

Check service logs (if something doesn't work):

docker compose -p localai -f docker-compose.local.yml logs n8n

Replace n8n with any service name

💾 Where is my data stored?

All data is in Docker volumes:

List all volumes:

docker volume ls | grep localai_

Your important data:

• localai_n8n_storage - All your workflows
• localai_langfuse_postgres_data - All databases
• localai_ollama_storage - Downloaded AI models

Backup data:

# During uninstall: Select "Yes" for backup
# Manual backup: Run the uninstall script and it creates automatic backup

Where are backups:

• Location: ~/ai-launchkit-backup-YYYYMMDD-HHMMSS/
• Contains: workflows, databases, all volumes as .tar.gz files

🆘 I'm stuck! Where can I get help?

Don't worry - help is available!

1. Check the logs:

cd ai-launchkit-local
docker compose -p localai -f docker-compose.local.yml logs

2. Check specific service:

docker logs n8n
docker logs flowise
docker logs postgres

3. Ask for help:

• GitHub Issues: Report a problem

  • Include: Error message, what you were doing, Ubuntu version

• Original AI LaunchKit: Main project

• Community Forum: oTTomator Think Tank

4. Common commands for troubleshooting:

# Check if Docker is running
sudo systemctl status docker

# Check system resources
htop  # Press Q to exit

# Check disk space
df -h

# Check memory
free -h

# Restart Docker
sudo systemctl restart docker

---

🌐 Service Port Schema

All services are accessible via http://SERVER_IP:PORT:

Core Services (8000-8019)

Port	Service	Description
8000	n8n	Workflow Automation Platform
8001	PostgreSQL	Database (external access)
8002	Redis	Cache Database (external access)
8003	Grafana	Monitoring Dashboards
8004	Prometheus	Metrics Collection
8005	Node Exporter	System Metrics
8006	cAdvisor	Container Monitoring
8007	Portainer	Docker Management UI

AI Services (8020-8039)

Port	Service	Description
8020	Open WebUI	ChatGPT-like Interface
8021	Ollama	Local LLM Runtime
8022	Flowise	AI Agent Builder
8023	bolt.diy	AI Web Development
8024	ComfyUI	Image Generation
8025	OpenUI	AI UI Component Generator
8026	Qdrant	Vector Database
8027	Weaviate	Vector Database with API
8028	Neo4j	Graph Database
8029	LightRAG	Graph-based RAG
8030	RAGApp	RAG Interface
8031	Letta	Agent Server

Learning Tools (8040-8050)

Port	Service	Description	Setup Guide
8040	Cal.com	Scheduling Platform	-
8047	Baserow	Airtable Alternative	→ Setup Guide
8048	NocoDB	Smart Spreadsheet	-
8049	Vikunja	Task Management	-
8050	Leantime	Project Management	-

Utilities (8060-8079)

Port	Service	Description	Setup Guide
8060	Postiz	Social Media Manager	→ Providers Setup
8062	Kopia	Backup System	-

Mail Services (8071 only - Learning/Testing)

Port	Service	Description
8071	Mailpit Web UI	Email catcher for development & testing

Note: Mailpit captures ALL emails for learning purposes. No external email delivery.

Specialized Services (8080-8099)

Port	Service	Description	Setup Guide
8080	Whisper	Speech-to-Text	-
8081	OpenedAI-Speech	Text-to-Speech	-
8082	LibreTranslate	Translation Service	-
8083	Scriberr	Audio Transcription	→ Troubleshooting
8084	Tesseract OCR	Text Recognition (Fast)
8085	EasyOCR	Text Recognition (Quality)
8086	Stirling-PDF	PDF Tools Suite
8087	Chatterbox TTS	Advanced Text-to-Speech
8088	Chatterbox UI	TTS Web Interface
8089	SearXNG	Private Search Engine
8090	Perplexica	AI Search Engine
8091	Formbricks	Survey Platform
8092	Metabase	Business Intelligence
8093	Crawl4AI	Web Crawler
8094	Gotenberg	Document Conversion
8095	Python Runner	Custom Python Scripts

AI Observability Stack (8096-8099)

Port	Service	Description	Setup Guide
8096	Langfuse	LLM Performance Tracking	→ Integration Guide
8097	ClickHouse	Analytics Database	-
8098	MinIO	Object Storage	-
8099	MinIO Console	Storage Management	-

Research & Notebooks (8100-8110)

Port	Service	Description	Setup Guide
8100	Open Notebook	NotebookLM Alternative - Research Assistant	→ Setup
8101	Open Notebook API	REST API for Open Notebook	→ TTS Integration

Special Ports

Port	Service	Protocol	Description
7687	Neo4j Bolt	TCP	Graph Database Protocol

---

🎯 Use Cases

This AI LaunchKit serves multiple purposes - from education to professional deployment. Here are practical scenarios for each service category:

💼 Professional Use Cases

Team AI Infrastructure:

• Deploy private AI services for your organization
• No data leaves your network
• Full control over models and data
• Comply with data protection regulations

Development Environment:

• Build and test AI applications locally
• Prototype before cloud deployment
• Debug workflows in safe environment
• Test different models and configurations

Business Automation:

• Automate internal processes with n8n
• Build custom AI tools for your team
• Create private knowledge bases with RAG
• Process documents without external APIs

🎓 Learning & Education Scenarios

⭐ Open Notebook - Transform Content into AI Podcasts

NotebookLM Alternative - Fully Local & Private!

Open Notebook is the standout feature for creating professional AI-powered podcasts and research:

🎙️ Create Professional Podcasts From Any Content

• YouTube Videos → Multi-speaker AI podcast discussions
• PDFs & Documents → Engaging audio summaries with analysis
• Web Pages & Articles → Podcast episodes with AI hosts debating topics
• Audio/Video Files → Transcribed, analyzed, and converted to podcasts

🤖 Completely Local AI Integration

• Speech-to-Text: Faster Whisper with multi-language support
• Text-to-Speech: OpenedAI Speech + German Thorsten voice (native pronunciation!)
• LLM Processing: Ollama integration - works 100% offline
• Cloud Option: Also supports OpenAI, Anthropic, Groq (16+ providers)

🚀 Simple Workflow

1. Upload: Paste YouTube URL, upload PDF, or add audio file
2. Analyze: AI reads, understands, and structures content
3. Script: AI creates engaging multi-speaker podcast script
4. Generate: Choose 1-4 AI voices with different personalities
5. Download: Professional MP3 ready to publish

📖 Complete Documentation

• Setup Guide - Configuration and first steps
• TTS Integration - Speech services setup
• Full Feature Guide - Advanced features and examples

Perfect for: Content creators, educators, researchers, and anyone wanting to transform written/video content into engaging audio format!

---

🤖 AI Core Services

n8n - Workflow Automation Learning

• Beginner: Build your first "Hello World" workflow with 300+ templates
• Intermediate: Connect Ollama LLM to process incoming webhooks and auto-respond
• Advanced: Create multi-agent AI systems using tools, memory, and conditional logic

Ollama - Local LLM Experimentation

• Beginner: Run your first local AI model (qwen2.5:7b) and compare with GPT-4
• Intermediate: Test different models for specific tasks (coding, translation, analysis)
• Advanced: Fine-tune models and benchmark performance metrics

Flowise - AI Agent Builder

• Beginner: Build a chatbot using drag-and-drop nodes in 5 minutes
• Intermediate: Create a RAG system that searches your documents using Qdrant
• Advanced: Build autonomous agents with tool-calling and memory management

Open WebUI - Prompt Engineering Lab

• Beginner: Learn effective prompt engineering with instant feedback
• Intermediate: Compare different models side-by-side for the same prompts
• Advanced: Create custom model pipelines and share them with your team

🗄️ RAG & Vector Databases

Qdrant - Semantic Search Learning

• Beginner: Upload documents and perform your first vector similarity search
• Intermediate: Build a "Chat with your PDFs" application using n8n
• Advanced: Implement hybrid search combining keywords and semantic vectors

Weaviate - AI-Powered Recommendations

• Beginner: Import product data and get AI-generated recommendations
• Intermediate: Build a content recommendation engine with custom schemas
• Advanced: Implement multi-modal search across text, images, and metadata

LightRAG - Graph-Based Retrieval

• Beginner: Understand how knowledge graphs improve RAG accuracy
• Intermediate: Build a question-answering system with relationship awareness
• Advanced: Combine graph structure with vector embeddings for complex queries

Neo4j - Graph Database Mastery

• Beginner: Model real-world relationships (social networks, org charts)
• Intermediate: Write Cypher queries to find patterns in connected data
• Advanced: Build recommendation engines using graph algorithms

📚 Learning Tools

Cal.com - Scheduling Automation

• Beginner: Set up automated meeting scheduling with calendar sync
• Intermediate: Create custom booking workflows with n8n webhooks
• Advanced: Build AI-assisted meeting preparation with pre-call research

Baserow & NocoDB - No-Code Database Learning

• Beginner: Create your first database with forms and views in the browser
• Intermediate: Connect to n8n workflows for automated data processing
• Advanced: Build custom business applications with API integrations

Vikunja & Leantime - Project Management Workflows

• Beginner: Organize personal projects with Kanban boards and Gantt charts
• Intermediate: Automate task creation from emails using n8n + Mailpit
• Advanced: Build AI-powered project analysis and reporting systems

🎨 Specialized AI Services

ComfyUI - Image Generation Pipelines

• Beginner: Generate your first AI image using pre-built workflows
• Intermediate: Create custom node graphs for specific art styles
• Advanced: Build automated image processing pipelines with batch operations

bolt.diy - AI Coding Assistant

• Beginner: Generate a complete web app from a simple prompt
• Intermediate: Learn how AI assistants structure projects and write code
• Advanced: Compare Claude, GPT-4, and Groq for different coding tasks

Whisper + TTS - Voice AI Learning

• Beginner: Transcribe audio files and convert text back to speech
• Intermediate: Build voice-controlled workflows with n8n
• Advanced: Create real-time voice translation systems

OCR Bundle - Document Processing

• Beginner: Extract text from images and PDFs automatically
• Intermediate: Build automated invoice processing with n8n workflows
• Advanced: Compare Tesseract (fast) vs EasyOCR (accurate) for different document types

LibreTranslate - Translation Experiments

• Beginner: Translate text in 20+ languages without external APIs
• Intermediate: Build multilingual content workflows with n8n
• Advanced: Compare neural translation quality across different language pairs

Perplexica & SearXNG - Search Engine Learning

• Beginner: Understand privacy-focused search without tracking
• Intermediate: Build custom search APIs with filtering and ranking
• Advanced: Create AI-enhanced research workflows combining search + LLM analysis

🔗 Integration Learning Patterns

Pattern 1: n8n + Ollama + Qdrant Build a complete RAG system that:

1. Indexes documents into Qdrant
2. Retrieves relevant context on questions
3. Uses Ollama to generate informed answers

Pattern 2: Whisper + LLM + TTS Create a voice assistant that:

1. Transcribes speech with Whisper
2. Processes with local LLM
3. Responds with natural TTS

Pattern 3: Crawl4AI + LLM + Email Build a research assistant that:

1. Crawls websites on schedule
2. Summarizes content with LLM
3. Emails digests via Mailpit

Pattern 4: Cal.com + n8n + LLM Create smart scheduling that:

1. Receives booking webhooks
2. Analyzes meeting context with AI
3. Prepares briefing materials

📊 Learning Paths

Path 1: AI Automation Fundamentals (1-2 weeks)

1. Set up n8n + Ollama + Mailpit
2. Build 5 basic workflows with templates
3. Create your first AI-powered automation

Path 2: RAG System Development (2-4 weeks)

1. Learn vector databases with Qdrant
2. Build document ingestion pipelines
3. Create production-ready RAG applications

Path 3: Multi-Agent Systems (4-8 weeks)

1. Master Flowise agent building
2. Implement tool-calling and memory
3. Build autonomous multi-agent workflows

Path 4: Voice AI Development (2-3 weeks)

1. Learn transcription with Whisper
2. Process audio with LLMs
3. Generate natural speech responses

💡 Learning Tips

Start Small: Begin with 3-5 core services (n8n, Ollama, Flowise, Mailpit, Monitoring)

Progressive Complexity: Master one service before adding others

Documentation Everything: Use n8n's notes feature to document your learning

Experiment Safely: All services are isolated in Docker - break things and rebuild!

Monitor Performance: Use Grafana to understand resource usage patterns

Join Community: Share your learning projects in forums and Discord

---

🌐 Network Access

Automatic LAN Configuration (Default)

During installation, the wizard will:

1. ✅ Auto-detect your server's LAN IP (e.g., 192.168.1.100)
2. ✅ Configure SERVER_IP automatically in .env
3. ✅ Set up firewall rules for LAN access (ports 8000-8099)

After installation, services are immediately accessible from ANY device:

http://192.168.1.100:8000  # n8n from laptop
http://192.168.1.100:8022  # Flowise from phone
http://192.168.1.100:8003  # Grafana from tablet
http://192.168.1.100:8071  # Email interface from any device

No manual configuration needed! Just open the URL from any device on your network.

Localhost Only (Alternative)

If you declined LAN access during installation, services use localhost:

# Access only from server
http://127.0.0.1:8000  # n8n
http://127.0.0.1:8022  # Flowise  
http://127.0.0.1:8003  # Grafana

To enable LAN access later:

1. Find your LAN IP: ip addr show | grep 'inet ' | grep -v 127.0.0.1
2. Update .env: sed -i 's/SERVER_IP=127.0.0.1/SERVER_IP=192.168.1.100/' .env
3. Restart: docker compose -p localai -f docker-compose.local.yml restart
4. Add firewall rules: sudo ufw allow from 192.168.0.0/16 to any port 8000:8099

Firewall Status

Check your current firewall configuration:

# View firewall status
sudo ufw status

# If LAN access wasn't configured during installation, add it manually:
sudo ufw allow from 192.168.0.0/16 to any port 8000:8099
sudo ufw allow from 10.0.0.0/8 to any port 8000:8099
sudo ufw reload

---

ℹ️ How Configuration Works (Reference)

⚠️ IMPORTANT: Everything is configured automatically during installation! You do NOT need to do anything here - this section is just for reference.

📖 Click to see what the installation wizard configures automatically (Informational Only)

The installation wizard automatically handles all configuration:

What Gets Configured Automatically:

1. Environment File (.env)

• ✅ Created from .env.local.example template
• ✅ All passwords generated automatically (32+ characters)
• ✅ All settings configured by wizard
• ✅ You never need to edit .env manually during installation

2. Network Configuration

• ✅ SERVER_IP auto-detected (e.g., 192.168.1.100)
• ✅ Wizard asks you to confirm the detected IP
• ✅ Automatically written to .env
• ✅ All services configured to use this IP

3. Service Selection

• ✅ Interactive checkbox menu in wizard
• ✅ Use arrow keys to navigate
• ✅ Spacebar to select/deselect
• ✅ Automatically written to COMPOSE_PROFILES in .env

4. Ollama Hardware Selection

• ✅ Choose CPU, NVIDIA GPU, or AMD GPU in wizard
• ✅ NVIDIA Container Toolkit installed automatically if GPU selected
• ✅ Automatically configured in .env

5. Mail Configuration

• ✅ Mailpit configured automatically
• ✅ All services configured to use Mailpit
• ✅ No external mail server needed

6. Optional API Keys

• ✅ Wizard asks if you want to add OpenAI, Anthropic, Groq keys
• ✅ Press Enter to skip (can add later)
• ✅ Automatically written to .env if provided

After installation completes, everything is ready to use! No manual configuration needed.

---

🔧 Changing Configuration After Installation (Optional)

Only use this section if you want to make changes after installation.

📝 How to add or remove services

When to use this:

• You want to try additional services (e.g., add ComfyUI, bolt.diy)
• You want to remove services you're not using
• You want to enable GPU after starting with CPU

Step-by-Step:

1. Stop all services:

   cd ~/ai-launchkit-local
   docker compose -p localai -f docker-compose.local.yml down

   • This stops all containers safely
   • Your data is preserved in volumes

2. Open .env file:

   nano .env

   • nano is a simple text editor
   • Press arrow keys to navigate

3. Find the COMPOSE_PROFILES line:

   COMPOSE_PROFILES="n8n,flowise,monitoring"

   • It's near the bottom of the file
   • Lists all active services

4. Edit the services:

   • Add services: COMPOSE_PROFILES="n8n,flowise,monitoring,cpu,comfyui"
   • Remove services: Delete the service name
   • Separate with commas (no spaces!)

   Available services:

   • AI: n8n, flowise, bolt, openui, comfyui, cpu, gpu-nvidia, open-webui
   • RAG: qdrant, weaviate, neo4j, lightrag, ragapp
   • Learning: calcom, baserow, nocodb, vikunja, leantime
   • Tools: kopia, postiz, monitoring
   • Specialized: speech, ocr, libretranslate, stirling-pdf, searxng, perplexica

5. Save and exit:

   • Press Ctrl+X
   • Press Y (yes, save changes)
   • Press Enter

6. Start services with new configuration:

   docker compose -p localai -f docker-compose.local.yml up -d

   • Starts services with your new selection
   • Downloads new service images if needed
   • Takes 2-5 minutes

7. Verify services are running:

   docker ps

   • Lists all running containers
   • Check for your new services

8. Access new services:

   • Open browser: http://SERVER-IP/
   • Click on newly added services
   • May need to wait 1-2 minutes for initialization

🌐 How to change server IP address

When to use this:

• Your server got a new IP address
• You want to access from different network
• You initially skipped LAN configuration

Step-by-Step:

1. Find your server's current IP:

   ip addr show | grep 'inet ' | grep -v 127.0.0.1

   • Shows all network interfaces
   • Look for your LAN IP (e.g., 192.168.1.100)
   • Example output: inet 192.168.1.100/24

2. Stop all services:

   cd ~/ai-launchkit-local
   docker compose -p localai -f docker-compose.local.yml down

3. Edit .env file:

   nano .env

4. Find SERVER_IP line:

   SERVER_IP=192.168.1.100

   • Usually near the bottom of the file

5. Change to new IP:

   SERVER_IP=192.168.1.200  # Your new IP

6. Save and exit:

   • Press Ctrl+X, then Y, then Enter

7. Restart all services:

   docker compose -p localai -f docker-compose.local.yml up -d

8. Test access:

   • Open browser: http://NEW-IP:8000
   • Services should load with new IP
   • Update bookmarks on your devices

🔑 How to add AI API keys

When to use this:

• You skipped API keys during installation
• You got new API keys
• You want to use cloud AI services (OpenAI, Anthropic, Groq)

Step-by-Step:

1. Open .env file:

   cd ~/ai-launchkit-local
   nano .env

2. Find the API key section:

   # Optional AI API Keys
   OPENAI_API_KEY=
   ANTHROPIC_API_KEY=
   GROQ_API_KEY=

3. Add your keys:

   OPENAI_API_KEY=sk-your-key-here
   ANTHROPIC_API_KEY=sk-ant-your-key-here
   GROQ_API_KEY=gsk-your-key-here

   • Remove the trailing = and add your key
   • No quotes needed
   • Get keys from: OpenAI.com, Anthropic.com, Groq.com

4. Save and exit:

   • Press Ctrl+X, then Y, then Enter

5. Restart affected services:

   docker compose -p localai -f docker-compose.local.yml restart n8n flowise bolt

   • Only restarts services that use API keys
   • Faster than restarting everything

6. Test API keys:

   • Open n8n: http://SERVER-IP:8000
   • Create workflow with OpenAI node
   • API keys should work now

⚙️ How to change n8n worker count

When to use this:

• System has more CPU cores
• Want to process workflows faster in parallel
• Currently have performance issues

Step-by-Step:

1. Check current CPU cores:

   nproc

   • Shows number of CPU cores
   • Example output: 8

2. Edit .env file:

   nano .env

3. Find N8N_WORKER_COUNT:

   N8N_WORKER_COUNT=1

4. Change number:

   N8N_WORKER_COUNT=4  # Use 4 workers

   • Recommendation: Use 50% of CPU cores
   • Don't exceed number of cores

5. Save and exit:

   • Press Ctrl+X, then Y, then Enter

6. Restart n8n:

   docker compose -p localai -f docker-compose.local.yml restart n8n-worker

7. Verify workers:

   docker ps | grep n8n-worker

   • Should show multiple n8n-worker containers
   • One per worker count

---

🗑️ Uninstalling AI LaunchKit

If you need to remove AI LaunchKit from your system:

Safe Uninstall with Backup

# Run the uninstall script
sudo bash ./scripts/uninstall_local.sh

The uninstall script will:

1. ✅ Show current AI LaunchKit status
2. ✅ Ask for confirmation before proceeding
3. ✅ Offer to create backup (workflows, databases, volumes)
4. ✅ Remove only AI LaunchKit containers and volumes
5. ✅ Preserve Portainer (or install it if missing)
6. ✅ Optionally keep or remove .env configuration

Manual Uninstall

If you prefer manual removal:

# Stop all services
docker compose -p localai -f docker-compose.local.yml down

# Remove with volumes (⚠️ DATA LOSS!)
docker compose -p localai -f docker-compose.local.yml down -v

# Remove images (optional)
docker image prune -a -f --filter "label=com.docker.compose.project=localai"

What Gets Removed

• ❌ All AI LaunchKit containers (n8n, Flowise, Ollama, etc.)
• ❌ All data volumes (workflows, databases, uploaded files)
• ❌ AI LaunchKit Docker networks
• ❌ Unused AI LaunchKit Docker images

What Gets Preserved

• ✅ Portainer (Docker Management UI)
• ✅ Other Docker containers not part of AI LaunchKit
• ✅ Project directory and scripts (can reinstall anytime)
• ✅ Your .env configuration (optionally backed up)

---

🔧 Management Commands

Service Management

# Start all services
docker compose -p localai -f docker-compose.local.yml up -d

# Stop all services  
docker compose -p localai -f docker-compose.local.yml down

# Restart specific service
docker compose -p localai -f docker-compose.local.yml restart n8n

# View service logs
docker compose -p localai -f docker-compose.local.yml logs n8n

# Check running services
docker ps

# Monitor resources
docker stats

Service Health Checks

# Check all service health
./scripts/06_final_report_local.sh

# Test specific port
nc -z localhost 8000

# Check port usage
netstat -tuln | grep 80

Updates

Automatic Update (Recommended):

# Run the update script
cd ai-launchkit-local
sudo bash ./scripts/update_local.sh

The update script will:

1. ✅ Create automatic backup of your configuration
2. ✅ Pull latest changes from GitHub
3. ✅ Update all Docker images
4. ✅ Restart services with new versions
5. ✅ Perform health checks
6. ✅ Provide rollback instructions if needed

Manual Update:

cd ai-launchkit-local

# Update repository
git pull origin main

# Update Docker images
docker compose -p localai -f docker-compose.local.yml pull

# Restart services
docker compose -p localai -f docker-compose.local.yml up -d

# Clean up old images
docker image prune -f

What Gets Updated:

• ✅ AI LaunchKit scripts and configurations
• ✅ Docker images for all services
• ✅ Landing page and templates
• ✅ Documentation

What Gets Preserved:

• ✅ Your .env configuration (automatically backed up and restored)
• ✅ All data in Docker volumes
• ✅ Service selections and settings

---

📊 Service-Specific Configuration

n8n Workflow Automation

• Access: http://SERVER_IP:8000
• First login: Create admin account on first visit
• Workflows: 300+ templates can be imported during installation
• API: Internal services use http://n8n:5678

Grafana Monitoring

• Access: http://SERVER_IP:8003
• Login: admin / [Check GRAFANA_ADMIN_PASSWORD in .env]
• Dashboards: Pre-configured for Docker monitoring
• Data Sources: Prometheus, PostgreSQL

Mailpit Email Testing

• Web UI: http://SERVER_IP:8071
• SMTP: SERVER_IP:8070 (port 1025 internal)
• Purpose: Captures all outgoing emails from services
• No Auth: Open access for local network

Database Access

• PostgreSQL: SERVER_IP:8001
• Username: postgres
• Password: Check POSTGRES_PASSWORD in .env
• Databases: Multiple apps share this instance

---

🔐 Security for Local Networks

Network Isolation

• Services only accessible from local network
• No external SSL certificates
• No internet-facing endpoints
• Docker network isolation between services

Authentication

• Disabled by default for local network convenience
• Each service has its own user management system
• Passwords stored in .env file
• No Basic Auth layers (unlike original)

Production Considerations

If deploying to production network:

1. Enable authentication on individual services
2. Configure SSL termination (nginx/Apache)
3. Restrict network access with firewall rules
4. Use strong passwords and API keys
5. Consider VPN access for remote users

---

🌐 Network Access Examples

From Server (Localhost)

# n8n Workflow Automation
curl http://127.0.0.1:8000

# Flowise AI Agent Builder  
curl http://127.0.0.1:8022

# Grafana Monitoring
curl http://127.0.0.1:8003

From Network Device

# Replace 192.168.1.100 with your server's IP
curl http://192.168.1.100:8000  # n8n
curl http://192.168.1.100:8022  # Flowise
curl http://192.168.1.100:8003  # Grafana

# From browser on phone/laptop/tablet
http://192.168.1.100/         # Service Dashboard
http://192.168.1.100:8000     # n8n interface
http://192.168.1.100:8071     # Email interface

API Integration

// n8n webhook from external service
POST http://192.168.1.100:8000/webhook/your-webhook-id

// Ollama API call
POST http://192.168.1.100:8021/api/generate
{
  "model": "qwen2.5:7b-instruct-q4_K_M", 
  "prompt": "Hello world"
}

// Vector search with Qdrant
POST http://192.168.1.100:8026/collections/search

---

📧 Mail System

Mailpit (Always Active)

• Purpose: Captures ALL emails sent by any service
• Web Interface: http://SERVER_IP:8071
• SMTP Server: SERVER_IP:8070 (internal port 1025)
• Authentication: None needed (local network)
• Storage: Emails stored in Docker volume (mailpit_data)

Configuration for Services

All services automatically use Mailpit:

SMTP_HOST=mailpit
SMTP_PORT=1025  
SMTP_USER=admin
SMTP_PASS=admin
SMTP_SECURE=false

Testing Email

1. Open any service that sends emails (n8n, Cal.com, etc.)
2. Trigger an email action
3. Check Mailpit web interface: http://SERVER_IP:8071
4. View email content, headers, attachments

---

🔧 Troubleshooting

Services Not Starting

Check port conflicts:

netstat -tuln | grep 80
# Look for ports in range 8000-8099

Check Docker resources:

docker stats
free -h
df -h

View service logs:

docker compose -p localai -f docker-compose.local.yml logs [service_name]

Network Access Issues

Can't access from other devices:

1. Check SERVER_IP in .env matches server's LAN IP
2. Verify firewall allows access:

   sudo ufw status
   sudo ufw allow from 192.168.1.0/24 to any port 8000:8099

3. Test connectivity: telnet SERVER_IP 8000

Services returning 404/502:

1. Wait 2-3 minutes for services to fully start
2. Check service is running: docker ps | grep service_name
3. Check port binding: docker port container_name

Database Connection Issues

Services can't connect to PostgreSQL:

# Check PostgreSQL is running
docker ps | grep postgres

# Test connection from service
docker exec n8n nc -zv postgres 5432

# Check logs
docker logs postgres

Performance Issues

High memory usage:

# Reduce n8n workers
echo "N8N_WORKER_COUNT=1" >> .env
docker compose -p localai -f docker-compose.local.yml restart

# Disable resource-heavy services temporarily
docker compose -p localai -f docker-compose.local.yml stop comfyui langfuse-web

Slow response times:

• Check available RAM: free -h
• Monitor CPU: htop
• Add swap: sudo fallocate -l 4G /swapfile && sudo swapon /swapfile

Common Service Issues

n8n Not Accessible

# Check n8n container status
docker logs n8n --tail 50

# Check database connection
docker exec n8n nc -zv postgres 5432

# Restart n8n
docker compose -p localai -f docker-compose.local.yml restart n8n

Common causes:

• Database not ready (wait 2-3 minutes)
• Workflow import still running (check docker logs n8n-import)
• Port 8000 already in use

Flowise Not Loading

# Check Flowise logs
docker logs flowise --tail 50

# Verify port binding
docker port flowise

# Test direct access
curl http://localhost:8022

Common causes:

• Container still initializing (wait 1-2 minutes)
• Port conflict on 8022
• Missing environment variables

Email Not Working

# Check Mailpit is running
docker ps | grep mailpit

# Test SMTP connection
docker exec n8n nc -zv mailpit 1025

# Check Mailpit logs
docker logs mailpit

Email test:

1. Open n8n: http://SERVER_IP:8000
2. Create simple workflow: Manual Trigger → Send Email
3. Execute workflow
4. Check emails: http://SERVER_IP:8071

---

📈 Performance Optimization

Resource Requirements by Service Count

Minimal (n8n + Flowise + Monitoring):

• RAM: 4GB
• CPU: 2 cores
• Services: ~8 containers

Standard (+ Business Tools):

• RAM: 8GB
• CPU: 4 cores
• Services: ~15 containers

Full Stack (All Services):

• RAM: 16GB+
• CPU: 8 cores
• Services: 40+ containers

Performance Tuning

Reduce n8n workers:

echo "N8N_WORKER_COUNT=1" >> .env

Optimize Baserow:

# Already configured in docker-compose:
BASEROW_RUN_MINIMAL=yes
BASEROW_AMOUNT_OF_WORKERS=1

Limit LibreTranslate models:

echo "LIBRETRANSLATE_LOAD_ONLY=en,de,fr" >> .env

Disable telemetry: Services have telemetry disabled by default for privacy and performance.

---

🔄 Migration from Domain-based Installation

If you have an existing domain-based AI LaunchKit installation:

Backup Data

# Export n8n workflows
docker exec n8n n8n export:workflow --backup --output=/backup/workflows.json

# Backup databases
docker exec postgres pg_dumpall -U postgres > backup.sql

# Backup volumes
docker run --rm -v localai_n8n_storage:/data -v $(pwd):/backup alpine \
  tar czf /backup/n8n_backup.tar.gz /data

Migration Steps

1. Deploy this local version on a new server
2. Import workflows via n8n interface
3. Restore database data if needed
4. Update any hardcoded URLs in workflows
5. Test all integrations with new IP:PORT format

---

🆘 Support

Documentation

• Original Project: AI LaunchKit
• Local Version: This README
• Service URLs: Generated LOCAL_ACCESS_URLS.txt

Getting Help

1. Check Logs: docker compose logs [service_name]
2. Service Status: docker ps
3. Resource Usage: docker stats
4. Port Conflicts: netstat -tuln | grep 80

Reporting Issues

• GitHub: Report local network issues
• Original: AI LaunchKit issues

Community

• Forum: oTTomator Think Tank
• Discord: Join the AI development community

---

📁 File Structure

ai-launchkit-local/
├── docker-compose.local.yml        # Port-based service definitions
├── .env.local.example              # Local network configuration template
├── scripts/
│   ├── install_local.sh            # Main installation script
│   ├── 01_system_preparation.sh    # System setup & firewall
│   ├── 02_install_docker.sh        # Docker installation
│   ├── 02a_install_nvidia_toolkit.sh # NVIDIA GPU support
│   ├── 03_generate_secrets_local.sh # Password generation
│   ├── 04_wizard_local.sh          # Interactive service selection
│   ├── 04a_setup_perplexica.sh     # Perplexica AI search setup
│   ├── 04b_setup_german_voice.sh   # German TTS voice auto-install
│   ├── 05_run_services_local.sh    # Service startup
│   ├── 06_final_report_local.sh    # Installation summary
│   ├── update_local.sh             # Update all services
│   ├── uninstall_local.sh          # Safe uninstall with backup
│   ├── generate_landing_page.sh    # Generate service dashboard
│   └── utils.sh                    # Shared utility functions
├── docs/                           # Complete documentation
│   ├── OPEN_NOTEBOOK_GUIDE.md      # Open Notebook features & use cases
│   ├── OPEN_NOTEBOOK_SETUP.md      # Open Notebook configuration
│   ├── OPEN_NOTEBOOK_TTS_INTEGRATION.md # Speech services setup
│   ├── QDRANT_SETUP.md             # Vector database API key setup
│   ├── GOTENBERG_GUIDE.md          # Document conversion API
│   ├── LANGFUSE_OLLAMA_INTEGRATION.md # LLM tracking setup
│   └── [service-specific guides]/  # Individual service documentation
├── templates/
│   ├── landing-page.html           # Auto-generated service dashboard
│   └── voice_to_speaker.yaml       # German TTS voice configuration
├── n8n/
│   ├── backup/workflows/           # 300+ pre-built n8n workflows
│   └── n8n_import_script.sh        # Workflow import automation
├── grafana/
│   ├── dashboards/                 # Pre-configured monitoring dashboards
│   └── provisioning/               # Auto-configuration for data sources
├── prometheus/
│   └── prometheus.yml              # Metrics collection configuration
├── shared/                         # Shared files between services
├── media/                          # Media processing workspace
└── [runtime directories]/          # Created during installation
    ├── open-notebook/              # Research assistant data
    ├── openedai-voices/            # TTS voice models
    ├── openedai-config/            # TTS configuration
    ├── perplexica/                 # AI search engine (git cloned)
    └── website/                    # Generated landing page

Note: Runtime directories are created automatically during installation and excluded via .gitignore.

---

🔄 Advanced Usage

Custom Service Configuration

Add custom ports:

# In docker-compose.local.yml
services:
  my-service:
    image: my-app:latest
    ports:
      - "8999:3000"  # Use ports outside main range

Custom environment:

# In .env
MY_SERVICE_CONFIG=value
MY_API_KEY=secret

Integration Examples

n8n workflow calling Ollama:

// HTTP Request Node in n8n
Method: POST
URL: http://ollama:11434/api/generate
Body: {
  "model": "qwen2.5:7b-instruct-q4_K_M",
  "prompt": "Hello from n8n workflow!"
}

External API calling services:

// From external application
const response = await fetch('http://192.168.1.100:8000/webhook/my-webhook', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({ message: 'Hello AI LaunchKit!' })
});

Multi-Server Setup

Deploy multiple instances:

# Server 1: Core AI services
COMPOSE_PROFILES="n8n,flowise,cpu,open-webui,monitoring"

# Server 2: RAG & Vector databases  
COMPOSE_PROFILES="qdrant,weaviate,neo4j,lightrag,ragapp"

# Server 3: Specialized services
COMPOSE_PROFILES="speech,ocr,libretranslate,stirling-pdf,comfyui"

---

📜 License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

Based on AI LaunchKit by Friedemann Schuetz.

---

Ready to launch your local AI stack?

💬 Issues & Feature Requests