Learning Hub

A self-hosted, single-tenant AI learning system that generates structured multi-lesson courses from user-defined topics using configurable LLM providers (Claude CLI, API, or local models). The system uses a bootstrap configuration flow and serves generated content as static HTML accessible from any device.

Quick Start (Docker)

git clone https://github.com/alexunder18/learning-hub.git
cd learning-hub
docker compose up -d

Open http://localhost:8080 - the browser-based setup will guide you through provider configuration.

Quick Start (Manual)

git clone https://github.com/alexunder18/learning-hub.git
cd learning-hub
python -m venv venv
.\venv\Scripts\activate
pip install -r requirements.txt
python server.py

Open http://localhost:8080

AI Providers

Provider	Cost	Requirements
Claude Code CLI	Included in Max/Pro subscription	Claude Code installed locally
Anthropic API	Pay per token	API key
OpenAI API	Pay per token	API key
Ollama	Free	Ollama running locally

See PROVIDERS.md for detailed configuration.

How It Works

1. Click + to create a new topic
2. Chat with AI about what you want to learn
3. Review and approve the proposed syllabus
4. Lessons are generated automatically
5. Access your lessons anytime on your home network

Configuration

On first run, the browser-based setup page configures your config.json. You can also create it manually:

{
  "mode": "cli",
  "provider": "claude-code",
  "model": "claude-opus-4-6",
  "api_key": "",
  "api_base": "",
  "initial_lessons": 3,
  "topics_dir": "./topics"
}

Config fields

Field	Values	Description
mode	cli, api	cli runs Claude Code CLI locally, api calls provider HTTP APIs
provider	claude-code, anthropic, openai, ollama	Which AI service to use
model	any model ID	e.g. claude-opus-4-6, claude-sonnet-4-6, gpt-4o, llama3
api_key	string	API key for Anthropic or OpenAI (not needed for CLI or Ollama)
api_base	URL	Custom API endpoint (only needed for Ollama, defaults to http://localhost:11434)
initial_lessons	number	How many lessons to generate when a new topic is created (default: 3)
topics_dir	path	Where topics are stored (default: ./topics)

Example configs

Claude Code CLI (requires local Claude Code installation + subscription):

{ "mode": "cli", "provider": "claude-code", "model": "claude-opus-4-6" }

Anthropic API (requires API key):

{ "mode": "api", "provider": "anthropic", "model": "claude-sonnet-4-6", "api_key": "sk-ant-..." }

OpenAI API (requires API key):

{ "mode": "api", "provider": "openai", "model": "gpt-4o", "api_key": "sk-..." }

Ollama (free, local, no account needed):

{ "mode": "api", "provider": "ollama", "model": "llama3", "api_base": "http://localhost:11434" }

Docker with Ollama (Free, Local)

# Uncomment the ollama service in docker-compose.yml, then:
docker compose up -d
docker compose exec ollama ollama pull llama3

Configure the setup page with provider "Ollama" and model "llama3".

Project Structure

├── server.py          # Flask application
├── config.json        # AI provider config (created on setup)
├── topics/            # Generated content (persisted)
├── templates/         # Jinja2 templates
├── static/            # CSS, JS
├── Dockerfile
└── docker-compose.yml