Local-first AI·AI / ML·2026

NeuroVault + Synapse

Your knowledge. A more useful memory.

In Development

Tech Stack

PythonOllamaRAGLocal LLMs

About

NeuroVault + Synapse is a local-first personal knowledge system. NeuroVault acts as the memory layer — ingesting, indexing, and retrieving documents using retrieval-augmented generation (RAG). Synapse is the reasoning layer — running local language models via Ollama to answer questions, surface connections, and assist with synthesis, entirely without sending data to external APIs.

The Problem

Cloud-based knowledge tools and AI assistants require uploading sensitive personal or research data to external servers, raising privacy concerns and introducing latency.

The Approach

By combining a local RAG pipeline with locally hosted language models via Ollama, NeuroVault + Synapse keeps all data and inference on the user's own machine.

Key Features

  • Local document ingestion and indexing
  • RAG-based retrieval from personal knowledge base
  • Local LLM inference via Ollama (no API key required)
  • Separate memory and reasoning layers
  • Privacy-first: no data leaves the device

Highlights

  • Designed separate knowledge and reasoning layers for retrieval-augmented workflows.
  • Local-first architecture that prioritizes private data and minimal exposure.