itsXactlY

itsXactlY / neural-memory

Public

Semantic memory system with knowledge graph, spreading activation, embedding-based recall, autonomous dream consolidation, and C++ LSTM+kNN pattern learning for the Hermes Agent.

agent agents ai avx2 avx2-simd
18
6
100% credibility
Found Apr 19, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A plugin for the Hermes AI agent that adds semantic long-term memory storage with automatic knowledge graph building, sleep-like consolidation phases, and pattern-learning enhancements for better recall.

How It Works

1
🔍 Discover smarter AI memory

You find a tool that helps your AI chat buddy remember conversations, facts, and preferences like a real friend, without forgetting important details.

2
📦 Set up the memory home

Run a couple of simple setup commands to create a safe place where your AI can store and organize all its memories.

3
🔗 Connect to your AI assistant

Tell your AI helper to use this new memory system by adding one line to its settings.

4
🧠 Watch your AI remember everything

Start chatting – now your AI recalls past talks, connects related ideas, and even 'dreams' to strengthen memories overnight.

5
📊 Peek at the memory web

Open a colorful dashboard to see how memories link together like a growing web of knowledge.

✨ Your AI feels alive

Enjoy conversations where your AI truly knows you, remembers details, and gets smarter over time.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 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 neural-memory?

Neural-memory is a Python-based semantic memory system for the Hermes Agent, storing facts and contexts as embeddings in a knowledge graph for similarity-based recall and spreading activation. It solves the problem of agents forgetting key details across sessions by adding autonomous "dream" consolidation that strengthens connections during idle time, plus C++-accelerated LSTM+kNN for predicting relevant memories from access patterns. Users get CLI tools like neural_remember, neural_recall, neural_think, and a live 3D dashboard, with SQLite lite mode or shared MSSQL for multi-agent setups.

Why is it gaining traction?

It stands out from basic key-value memory or github semantic kernel plugins by mimicking brain-like processes—NREM replay, REM bridging, insight clustering—in a neural memory network that's fully local and offline. Developers hook on the pattern learning that re-ranks results by temporal frequency and graph proximity, boosting recall accuracy without cloud APIs. Easy install scripts handle lite/full modes, migrations fix bloat, and tests verify production DB integrity.

Who should use this?

Hermes Agent builders needing persistent example semantic memory for user prefs, facts, or long conversations. AI researchers prototyping memory augmented neural networks, neural memory LLMs, or long short memory neural network agents. Devs tired of manual context injection in semantic github actions or semantic memory modules.

Verdict

Try it if you're on Hermes—solid user tools and dream engine deliver real memory smarts, despite low maturity (18 stars, 1.0% credibility). Docs and migrations are production-grade; run tests post-install to confirm.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.