Revornix

A self-hosted AI knowledge workspace that turns documents and audio into searchable Markdown, illustrated reports and podcasts with OpenAI-compatible providers.

Revornix is an AI knowledge workspace for people who collect articles, research papers and recordings and want to ask questions across their library. It connects saved material in a personal knowledge graph and uses those relationships to inform its answers.

Document collection: Import web pages, PDFs, Word, Excel, PowerPoint, plain text and audio into Markdown, with speaker separation for meeting recordings.
Search and document chat: Search your private library by text or meaning, and ask an assistant questions grounded in your documents and their knowledge graph.
Reports and podcasts: Generate summaries, illustrated content and two-voice podcasts from documents or collections, with automatic daily digests and email, push or in-app notifications.
Editing and sharing: Edit Markdown with tables, Mermaid diagrams and math, then keep collections private, share them with collaborators or publish them on the web.
Assistant integrations: Connect external tools through MCP or expose your library to MCP-aware assistants; a public API and Python SDK support automated collection.

The full workspace can run on your own infrastructure, and a hosted workspace is also available. It has a browser interface and an Electron desktop shell for macOS and Windows. It supports OpenAI-compatible providers, with independently selectable engines for document conversion, embeddings, summaries, podcasts and illustrations. Self-hosting gives you control of the documents and databases; where AI requests go depends on the providers you choose.