Analysis updated 2026-05-18
Build a personal knowledge base that remembers facts you tell it across many chat sessions.
Explore how a hybrid graph plus vector memory system works by running it locally for free.
Toggle between a self-hosted open source memory backend and a managed cloud version for the same app.
Erase all stored memory for a dataset to test a privacy-focused data deletion flow.
| jaybamroliya/mindvault | abhishek-kumar09/mern-live-js-june-2021 | abhishek-kumar09/z | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | HTML | HTML | HTML |
| Last pushed | — | 2021-09-13 | 2019-12-23 |
| Maintenance | — | Dormant | Dormant |
| Setup difficulty | moderate | moderate | easy |
| Complexity | 3/5 | 2/5 | 1/5 |
| Audience | developer | developer | general |
Figures from each repo's GitHub metadata at analysis time.
Requires a free Groq API key for the language model, everything else runs locally at no cost.
MindVault is a personal memory app for AI, built during a hackathon for Cognee, an open source memory tool that combines a knowledge graph with vector search. The idea is that normal AI chats forget everything once the conversation ends, while MindVault builds a lasting, searchable memory of what you tell it, made up of facts and how those facts connect to each other, and it keeps that memory across as many sessions as you like. The app offers four core actions built on Cognee's memory system. Remember stores new text into permanent memory. Recall lets you ask questions and get answers pulled from both semantic search and graph connections. Improve reprocesses stored memories to strengthen the connections between them. Forget erases memory entirely, which the README describes as useful for meeting data privacy rules like GDPR. On top of that, MindVault shows a live, growing knowledge graph in the browser using plain JavaScript with no extra libraries, and it supports voice input through the browser's built in speech recognition. You can switch between running Cognee locally, which is free and open source, or using Cognee's managed cloud service, with a single toggle in the interface and no restart needed. Memories can also be organized into separate named groups, such as work, personal, or research. The backend is written in Python using FastAPI, with a file called memory_engine.py acting as a bridge between the local and cloud versions of Cognee so the rest of the app does not need to know which one is active. To run it locally, you clone the repository, install the Python dependencies, copy the example environment file, add a free Groq API key for the language model, and start the server with uvicorn. The README states the local setup needs no paid services at all, relying on Groq's free tier and a local embedding model called fastembed. MindVault is aimed at developers curious about combining graph and vector based memory for AI applications, or anyone who wants a personal, self hosted alternative to AI chat tools that forget everything between sessions.
A personal AI memory app that stores what you tell it in a knowledge graph, with local or cloud modes and a live graph view.
Mainly HTML. The stack also includes Python, FastAPI, Cognee.
No license information was included in the README.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.