IceXWiKi

View on GitHub

IceXWiKi

A self-hosted knowledge management platform combining XWiki with AI-powered features, infrastructure auto-discovery, GitHub integration, and Confluence migration.

Built as a Docker Compose stack with 5 services, designed for homelab environments.

XWiki Home

Features

  • XWiki Wiki Engine - Full-featured wiki with PostgreSQL backend
  • Bridge API - FastAPI middleware connecting XWiki to GitHub, Ollama AI, Word import, and AnythingLLM RAG
  • AutoDoc - Automated infrastructure discovery (Docker, Network, ESXi, Synology) with AI-annotated reports
  • Confluence Migration - Bulk migrate all Confluence spaces including attachments and content conversion
  • RAG Integration - Ingest wiki content into AnythingLLM for semantic search

Architecture

The stack consists of 5 Docker containers:

Service Description Port
XWiki Wiki engine (LTS, PostgreSQL backend) 8085
PostgreSQL Database backend 5433
Bridge FastAPI - GitHub sync, AI endpoints, Word import, RAG 8090
AutoDoc FastAPI - Infrastructure scanning + Web GUI 8091
AnythingLLM RAG engine with Ollama LLM/embedding provider 3001

External dependency: Ollama running on a separate host with qwen2.5:14b and nomic-embed-text models.

Screenshots

Bridge API (Swagger UI)

Bridge Swagger

AutoDoc Web GUI

AutoDoc GUI

AutoDoc API (Swagger UI)

AutoDoc Swagger

Quick Start 

git clone https://github.com/icepaule/IceXWiKi.git
cd IceXWiKi

# Configure environment
cp .env.example .env
# Edit .env with your actual values (passwords, IPs, tokens)

# Build and start
make setup
# or: docker compose up -d --build

# Wait ~2 minutes for XWiki initialization, then open:
# XWiki:    http://YOUR_HOST:8085
# Bridge:   http://YOUR_HOST:8090/docs
# AutoDoc:  http://YOUR_HOST:8091

Configuration

All configuration is via environment variables in .env (never committed). Copy .env.example and fill in your values:

Variable Description Example
XWIKI_PORT XWiki web port 8085
XWIKI_ADMIN_USER XWiki admin username admin
XWIKI_ADMIN_PASSWORD XWiki admin password CHANGEME
OLLAMA_URL Ollama API endpoint http://ollama-host:11434
OLLAMA_MODEL LLM model for AI features qwen2.5:14b
DOCKER_HOSTS Remote Docker TCP endpoints tcp://nas:2375
SCAN_SUBNETS Subnets for network scanning 192.168.1.0/24
GITHUB_USER GitHub username for repo sync your-user
GITHUB_TOKEN GitHub personal access token (optional, for private repos)
CONFLUENCE_URL Confluence server URL http://confluence:8090

See .env.example for all available options.

Services in Detail

Bridge API (Port 8090)

FastAPI service connecting XWiki to external systems.

Endpoints:

Endpoint Method Description
/api/github/sync POST Sync GitHub repos to XWiki pages
/api/import/word POST Import DOCX files as wiki pages
/api/ai/summarize POST AI-powered text summarization
/api/ai/runbook POST Generate runbooks from notes
/api/ai/classify POST Classify content categories
/api/rag/ingest-space POST Ingest XWiki space into AnythingLLM

AutoDoc (Port 8091)

Infrastructure auto-discovery service with web GUI and scheduled scanning.

Scanners:

Scanner Method Description
Docker Docker API (socket + TCP) Containers, images, networks, volumes
Network nmap Subnet discovery, service detection
ESXi SSH (paramiko, RSA key) VMs, datastores, networking
Synology SSH (paramiko, ed25519) Volumes, shares, packages

Features:

  • Web GUI for manual scans and configuration
  • APScheduler for periodic automated scans
  • AI analysis of scan results via Ollama
  • Results written as XWiki pages under AutoDoc/ space

Confluence Migration

Standalone Python script to migrate Confluence spaces to XWiki.

# Migrate all spaces
export $(grep -v '^#' .env | xargs)
python3 scripts/migrate_confluence.py --space ALL

# Dry run for a single space
python3 scripts/migrate_confluence.py --space NETOPS --dry-run

Capabilities:

  • Pagination support for large spaces
  • Confluence Storage Format (XHTML) to XWiki 2.1 syntax conversion
  • Attachment migration with content-type preservation
  • Macro mapping: tip -> success, note -> warning, info/warning preserved
  • Handles headings, bold, italic, code blocks, tables, links, images

AnythingLLM (Port 3001)

RAG engine for semantic search across wiki content. Configured to use Ollama as both LLM and embedding provider.

Project Structure 

IceXWiKi/
├── docker-compose.yml        # 5 services orchestration
├── .env.example              # Configuration template
├── Makefile                  # Shortcuts (up, down, sync, scan)
├── bridge/                   # Bridge FastAPI service
│   ├── Dockerfile
│   ├── requirements.txt
│   └── app/
│       ├── main.py           # App + health + router includes
│       ├── config.py         # Pydantic settings from ENV
│       ├── routers/          # API endpoints
│       │   ├── github_sync.py
│       │   ├── word_import.py
│       │   ├── ai_endpoints.py
│       │   └── anythingllm.py
│       └── services/         # External client libraries
│           ├── xwiki_client.py
│           ├── ollama_client.py
│           ├── github_client.py
│           └── anythingllm_client.py
├── autodoc/                  # AutoDoc FastAPI service
│   ├── Dockerfile
│   ├── requirements.txt
│   └── app/
│       ├── main.py           # App + GUI + scan endpoints
│       ├── config.py
│       ├── scheduler.py      # APScheduler periodic jobs
│       ├── scanners/         # Infrastructure scanners
│       │   ├── docker_scanner.py
│       │   ├── network_scanner.py
│       │   ├── esxi_scanner.py
│       │   └── synology_scanner.py
│       ├── services/
│       │   ├── xwiki_writer.py
│       │   └── ollama_analyzer.py
│       └── static/
│           └── index.html    # AutoDoc Web GUI
├── scripts/
│   ├── setup.sh              # Initial setup script
│   ├── migrate_confluence.py # Confluence migration
│   └── fix_macros.py         # Fix unsupported macros
└── docs/
    └── screenshots/          # Documentation images

Makefile Targets 

make up          # Start all services
make down        # Stop all services
make restart     # Restart all services
make logs        # Follow logs
make build       # Build custom images
make ps          # Show container status
make github-sync # Trigger GitHub repo sync
make scan-all    # Run all AutoDoc scans
make scan-docker # Run Docker scan only
make scan-network # Run network scan only
make clean       # Remove containers and volumes

Requirements

  • Docker Engine 24+ with Compose v2
  • Ollama instance with qwen2.5:14b and nomic-embed-text models
  • ~4 GB RAM for the stack
  • ~10 GB disk for XWiki data, PostgreSQL, and AnythingLLM

Optional:

  • SSH key access to ESXi and Synology hosts (for AutoDoc scanning)
  • Docker TCP API enabled on remote hosts (for remote container scanning)
  • Confluence server (for migration)
  • GitHub account (for repo sync)

Security Notes

  • No secrets in this repository - all credentials are in .env (gitignored)
  • .env.example contains only placeholder values
  • XWiki REST API uses Basic Auth
  • SSH scanning uses key-based authentication (keys mounted as volumes)
  • Docker socket is mounted read-only

License

MIT