🤖 ai-stack-wizard

Generate a complete local AI stack in the browser — private, powerful, no cloud required.

Pick your services · Choose your models · Download a ready-to-use docker-compose.yml

---

🎯 What is this?

Running a local AI stack means juggling Ollama, Open WebUI, n8n, vector databases, and more — before a single container is running. Finding the right images, configuring ports, wiring services together, and deciding which models fit your hardware takes hours.

ai-stack-wizard solves exactly that. Pick the services you need, select your Ollama models, enter your RAM to get tailored recommendations — and download a fully wired docker-compose.yml in seconds. No installation, no build system, no backend — a single HTML file.

---

✨ Features

🧠 RAM-aware model picker	Select your RAM and get instant model recommendations
✅ 12 pre-configured AI services	Ollama, n8n, Open WebUI, Flowise, LiteLLM and more
✅ Ollama model auto-pull	Selected models are pulled automatically on container start
✅ Pre-wired networking	All services communicate via a shared ai-stack Docker network
✅ Live YAML preview	Syntax highlighting directly in the browser
✅ One-click download	Finished docker-compose.yml ready to deploy
✅ No backend required	A single HTML file — open locally or host on GitHub Pages
✅ Fully configurable	Ports, paths, passwords, API keys — everything adjustable

---

📦 Included Services

🦙 Core Inference — Ollama

Service	Description
Ollama	Run large language models locally with a simple REST API. Models are pulled automatically on startup.

💬 Chat Interface — Open WebUI, AnythingLLM

Service	Description
Open WebUI	ChatGPT-like interface for your local Ollama models
AnythingLLM	All-in-one: chat, RAG pipelines, AI agents & workspaces

⚡ Workflow Automation — n8n

Service	Description
n8n	Automate anything with AI-powered workflows. Pre-wired to talk to Ollama directly.

🧠 AI Agents & RAG — Flowise, Qdrant, SearXNG

Service	Description
Flowise	Build LLM apps, RAG pipelines & AI agents with a visual editor
Qdrant	High-performance vector database for semantic search & RAG
SearXNG	Private, self-hosted web search engine for AI agents — no tracking

🔀 AI Gateway — LiteLLM

Service	Description
LiteLLM	Unified OpenAI-compatible API proxy for all your models — local and remote

📊 Observability — Langfuse, Komodo

Service	Description
Langfuse	LLM monitoring, tracing, prompt management & analytics
Komodo	Modern container & stack management UI (Portainer alternative)

🎙️ Audio & Speech — faster-whisper

Service	Description
faster-whisper	Local speech-to-text transcription API — meetings, voice notes, podcasts

📄 Document AI — Paperless-ngx

Service	Description
Paperless-ngx	Document management with OCR and full-text search — wired to local AI

---

🧠 Ollama Model Guide

Model	RAM	Best for
Llama 3.2 3B	~8 GB	Fast, lightweight everyday tasks
Mistral 7B	~8 GB	Excellent reasoning, very efficient
Llama 3.2 8B	~16 GB	Best balance of quality and speed
Gemma 3 9B	~16 GB	Coding & structured tasks
Qwen 2.5 14B	~16 GB	Multilingual & code generation
DeepSeek R1 8B	~16 GB	Math & complex reasoning
Phi-4 14B	~16 GB	Microsoft's compact powerhouse
Llama 3.1 70B	~48 GB	Near-GPT-4 quality
Nomic Embed	~4 GB	Text embeddings for RAG pipelines
LLaVA 13B	~16 GB	Vision — describe and analyze images

---

🚀 Getting Started

Option 1 — Open directly in the browser

# Clone the repository
git clone https://github.com/bitalchemy-io/ai-stack-wizard
cd ai-stack-wizard

# Open index.html in your browser
open index.html         # macOS
xdg-open index.html     # Linux

Option 2 — Local dev server

python3 -m http.server 8080
# → http://localhost:8080

Option 3 — GitHub Pages (recommended)

git init
git add index.html README.md
git commit -m "initial commit"
git remote add origin https://github.com/bitalchemy-io/ai-stack-wizard
git push -u origin main

Then in GitHub: Settings → Pages → Branch: main → Save

The app will be available at https://bitalchemy-io.github.io/ai-stack-wizard — free, no server needed.

---

▶️ Running your stack

# Start everything
docker compose up -d

# Check status
docker compose ps

# Follow Ollama logs (model download progress)
docker compose logs -f ollama

Default ports at a glance

Service	Port	URL
Ollama API	11434	http://localhost:11434
Open WebUI	3000	http://localhost:3000
n8n	5678	http://localhost:5678
Flowise	3001	http://localhost:3001
LiteLLM	4000	http://localhost:4000
Qdrant	6333	http://localhost:6333
SearXNG	8888	http://localhost:8888
Langfuse	3002	http://localhost:3002
Whisper	8000	http://localhost:8000
Paperless-ngx	8010	http://localhost:8010
AnythingLLM	3010	http://localhost:3010
Komodo	9120	http://localhost:9120

---

🔧 Adding a Custom Service

Add a new entry to the SERVICES object in index.html:

my_service: {
  name: 'My Service',          // Display name
  icon: '🚀',                  // Emoji icon
  cat: 'AI Agents & RAG',      // Category
  desc: 'Short description',
  tag: 'image:tag',            // Docker image
  fields: [
    {
      key: 'port',
      label: 'Port',
      default: '8080',
      hint: 'Optional hint shown below the input'
    },
  ],
  yaml: (c) => `
  my-service:
    image: my-image:tag
    container_name: my_service
    restart: unless-stopped
    ports:
      - "${c.port}:8080"
    environment:
      - OLLAMA_BASE_URL=http://ollama:11434`.trim(),
  volumes: () => [],
},

---

🗺️ Roadmap

• GPU support detection (CUDA / Apple Silicon / ROCm)
• Generate .env file alongside the YAML
• Recommended stack presets (Minimal / RAG / Full)
• More services (Open Interpreter, Dify, PrivateGPT, LocalAI …)
• One-click litellm_config.yaml generation

---

🤝 Contributing

Pull requests are welcome! Especially looking for: new AI services, model updates, and improvements to existing YAML templates.

1. Fork the repository
2. Create a feature branch: git checkout -b feature/new-service
3. Commit your changes: git commit -m 'feat: add Dify'
4. Push the branch: git push origin feature/new-service
5. Open a Pull Request

---

📄 License

MIT — free to use, fork, and extend.

---

Made with ❤️ for the local AI community