AI Assistant
| AI Assistant | |
|---|---|
| Developer | Professional Wiki |
| Type | MediaWiki Extension |
| Initial release | 2025 |
| Operating system | Web (SaaS), Linux |
| Written in | PHP, JavaScript |
| License | Proprietary |
| Website | https://professional.wiki/ |
| Contents | |
AI Assistant is a proprietary MediaWiki extension developed by Professional Wiki that adds an AI-powered chatbot widget to any MediaWiki installation. It uses Retrieval-Augmented Generation (RAG) to answer questions grounded in the wiki's own content, providing answers with source attribution to specific wiki pages.
Key Features
- Embeddable chat widget that appears on wiki pages
- RAG architecture: answers are generated from the wiki's own content, not from general AI training data
- Source attribution: every answer links to the wiki pages used as sources
- Powered by the Anthropic Claude API
- Deployable in 2–3 days with no infrastructure changes required
- Works with any MediaWiki installation, including ProWiki and DockerWiki hosted wikis
- Keeps organizational knowledge sovereign — no wiki content is stored by the AI provider
How It Works
AI Assistant indexes the content of a MediaWiki wiki and stores it in a vector database. When a user asks a question via the chat widget, the system retrieves the most relevant wiki pages and sends them as context to a large language model. The model generates an answer based solely on that retrieved context, and the widget displays the answer alongside links to the source pages.
This approach — known as RAG (Retrieval-Augmented Generation) — ensures that the AI does not hallucinate information from outside the wiki and that answers remain accurate as wiki content is updated.
Enterprise Use
AI Assistant is designed for organizations that have invested in building a structured knowledge base and want to make that knowledge instantly accessible via natural language queries. Typical use cases include:
- Internal IT help desks and employee self-service portals
- Customer-facing support wikis and documentation sites
- Onboarding portals where new employees can ask questions about policies and processes
- Technical documentation search for engineering teams
The extension integrates with existing MediaWiki access controls, ensuring that users only receive answers based on content they are permitted to view.
Deployment
AI Assistant is sold as a proprietary extension at a fixed per-installation price. Professional Wiki handles the full deployment, including vector database setup, API key configuration, and widget integration. The typical deployment timeline is 2–3 days.
The extension can be deployed on:
- ProWiki managed hosting (SaaS)
- DockerWiki dedicated managed hosting
- Self-hosted MediaWiki installations
A live demo wiki with representative enterprise software content is available at test.pro.wiki, covering topics such as Knowledge Management, Artificial Intelligence, IT tools, and more.
