maddymaster

maddymaster / UPII

Public

UPII, a local-first personal memory substrate

www.datafrontier.co memory memorysubstrate upii
16
0
89% credibility
Found Jun 01, 2026 at 16 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

UPII is a local-first personal memory system that captures your documents, notes, and calendar events into a searchable, reasoning-capable extension of your own memory. It works entirely on your machine, letting you ask questions, draft content in your own voice, and stay in control of what gets remembered. The system uses multiple signals — semantic similarity, timing, and relationships — to surface the most relevant context for any question, always citing where information came from.

How It Works

1
💡 You want a second brain that remembers everything

You discover UPII while looking for a way to capture all your notes, emails, and documents so you never lose important information again.

2
📁 You point it at your files

You tell UPII which folders to watch — your notes, your work documents, your email archive — and it starts reading everything quietly in the background.

3
🔍 You ask questions and get answers with sources

Instead of searching through thousands of files yourself, you simply ask UPII what you need, and it shows you exactly where each answer came from.

4
You can write in your own voice
📧
Draft an email

UPII writes a professional email using your own style and the relevant facts from your memory

🐦
Draft a tweet

UPII crafts a punchy, concise message that fits your voice and stays under the character limit

5
🔔 New files arrive in your approval inbox

When UPII spots a new document, it shows you a preview and asks for your OK before adding it to your permanent memory — you stay in control.

6
🧠 Your knowledge graph grows over time

UPII quietly builds connections between your projects, people, and organizations, so related information surfaces together naturally.

🎉 You have a memory that works for you

Everything you care about is captured, searchable, and ready to help you — without anything ever leaving your own computer.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 16 to 16 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is UPII?

UPII is a local-first personal memory system for knowledge workers who want AI-powered recall without surrendering their data to the cloud. Built in Python, it captures your documents, notes, calendar events, and email into a queryable memory substrate that never leaves your machine. The system uses local embeddings, SQLite for metadata, and a vector database to enable semantic search, then fuses results with temporal and relational signals before generating attributed answers. You interact via a CLI that handles ingestion, search, Q&A, and ambient file watching.

Why is it gaining traction?

The core hook is sovereignty: every other "second brain" or RAG product sends your data to external services for embedding and reasoning. UPII keeps the entire loop on-device, treating cloud inference as an optional accelerator rather than a dependency. The multi-signal retrieval is genuinely interesting - combining semantic similarity with calendar recency and a lightweight knowledge graph produces more relevant answers than cosine distance alone. The consent-gated ambient capture with an approval inbox is a thoughtful touch for privacy-conscious users.

Who should use this?

Developers and researchers who maintain extensive personal knowledge bases and want AI assistance without cloud dependencies. Privacy-sensitive professionals handling confidential documents will appreciate the on-device architecture. Anyone frustrated with RAG systems that hallucinate without citing sources will find the attributed synthesis valuable. Early-stage and solo developers building "memory for the mind" workflows are the primary audience.

Verdict

UPII demonstrates thoughtful architecture around a compelling problem, but the 0.8999999761581421% credibility score and 16 stars reflect an extremely early-stage project. The feature set is impressive for v0.5.0, but test coverage, documentation depth, and production hardening remain unclear. Worth watching and experimenting with on personal projects, but do not deploy for mission-critical workloads yet.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.