Your documents.
Your local RAG.
A private, local-first desktop app that turns your files into an intelligent knowledge base using RAG. Powered by the LLM of your choice.
A private, local-first desktop app that turns your files into an intelligent knowledge base using RAG. Powered by the LLM of your choice.
Your data never leaves your device. Arivu runs entirely offline using local LLMs (Ollama) and embeddings.
Uses RAPTOR (Recursive Abstractive Processing) to summarize documents hierarchically for better context.
Switch between Similarity Search and MMR (Maximal Marginal Relevance) to balance relevance and diversity.
Customize everything. Swap embedding models, tweak chunk sizes, and experiment with different RAG strategies per project.
Integrated Tavily web search to augment your local knowledge base when the answer isn't in your docs.
Ingest PDFs, Word docs, Code files, Emails, and more. If it's text, Arivu can read it.
Drag and drop your files into a project workspace.
Local embeddings and RAPTOR generate a searchable knowledge base.
Ask questions and get answers with citations to your docs.
Since the app is not notarized by Apple, run this command in Terminal after moving the app to Applications:
xattr -rd com.apple.quarantine /Applications/Arivu.app
Give it a minute for the backend to load.