Revornix
Local-first AI workspace turning inputs into structured notes and podcasts.
Ingest messy streams, structure knowledge, and auto-generate two-voice podcasts.
Revornix is an open-source, local-first workspace that takes fragmented web links, documents, and audio inputs, organizes them via Milvus and Neo4j, and converts them into structured knowledge and automated two-voice podcasts. It suits audio producers and researchers who want to automate research digestion and generate spoken summaries without relying on locked-down cloud platforms.
What it is
The platform ingests multi-format inputs including audio files, transcribes them into searchable Markdown with speaker separation, and processes them through pluggable converters. It combines chunked vector storage with a personal GraphRAG layer to provide context-aware AI answers, automatic daily digests, and generated two-voice audio episodes.
Who it's for
It suits audio content producers, researchers, and creators drowning in unstructured information who want local control over their knowledge graphs and audio generation workflows. It does not suit casual note-takers who want a simple zero-config desktop app, as it requires managing multiple backing databases and a self-hosted stack.
Key features
Audio Transcription & Podcasts
Transcribes audio documents into searchable Markdown with speaker separation and turns content into two-voice audio episodes.
Personal GraphRAG
Projects saved documents onto a personal knowledge graph in Neo4j paired with chunked vector storage in Milvus.
Pluggable Converters
Cleans and normalizes web pages, PDFs, and office formats using engines like MinerU or Jina.
MCP Client and Server
Drives external Model Context Protocol servers and exposes your personal library to MCP-aware tools.
Trending Feed Aggregator
Gathers aggregated hot-search headlines across mainstream platforms via the bundled hot-news service.
Up and running in 3 steps
Our honest verdict
What's great
- Turns unstructured audio and documents into two-voice podcasts automatically
- Combines Milvus vector search with a personal Neo4j knowledge graph
- Supports pluggable document converters like MinerU and Jina
- Acts as both an MCP client and an MCP server for external tool integration
Where it falls short
- Requires spinning up a heavy local dependency bootstrap including Postgres, Redis, Neo4j, MinIO, and Milvus
- Self-hosted stack demands ongoing container management and infrastructure upkeep
- License is not stated in the repository metadata
Pricing & license
Free under its repository structure, with the only cost being the infrastructure and API keys required to run the self-hosted stack yourself.
Open remote AI-training roles
870 open remote AI-training roles, updated November 14, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Can Revornix handle audio files?
Yes, it supports audio ingestion and transcription into searchable Markdown with speaker separation, as well as generating two-voice podcast versions.
What databases are required to run it?
The local bootstrap relies on PostgreSQL, Redis, Neo4j, MinIO, and Milvus.
