morningmonkey.net
MORNING MONKEY

Where Ideas Wake Up

Morning Monkey is a launchpad for my personal projects. A web of infrastructure, code and data. Discover what I’m building next.

Morning Monkey
◆

LABS & INNOVATION

Where curiosity meets creation

Emanuel – The Loop Documentation Assistant

nsai.morningmonkey.net GitHub Repository
Status:
Working prototype under testing

The Challenge: The Loop Project is a miracle for Type 1 Diabetics, allowing the freedom to build a DIY artificial pancreas. The Loop project documentation is vast, technical, and complex. For many users, navigating the documentation feels overwhelming.

The Solution: Meet Emanuel, an AI powered documentation helpful assistant designed to make the Loop journey a little more manageable.

Emanuel Project Sketch
How It Works: Emanuel uses a RAG (Retrieval-Augmented Generation) architecture to scrape and index the entire Loop documentation library.
Instant Answers: Don't search for hours. "Is it ok to upgrade to IOS 26?" and get a sourced answer with a link to the documentation.
Language Bridge: Loop's docs are in English, but diabetes is global. Emanuel translates and interprets instructions into your native language instantly.

Insight – The AI Data Team for Loopers

nsai.morningmonkey.net GitHub Repository
Status:
Alpha development in progress, showing promising results. Testing hopefully during January 2026.

The Challenge: Living with a closed-loop system generates massive amounts of data. While Nightscout captures every heartbeat of the system, interpreting millions of data points stored in MongoDB is overwhelming. Most users can see what happened, but they struggle to understand why, or how to tune their system for better results.

The Solution: Meet The Insight Team. A suite of specialized AI agents that don't just read charts—they investigate them.

Insight Project Sketch
How It Works: Processing months of medical data is too heavy for an LLM alone. Insight uses a hybrid approach between classic data engineering and AI engineering. Business intelligence metrics are applied to the data to create a structured dataset that is then fed to the LLMs. This allows the LLMs to act as high-level analysts rather than raw calculators.
Secure Retrieval: Using google authentication the user profile is securely stored in our servers. Saving your nightscouts secure api token in your account will enable the system to fetch your data from your Nightscout instance.
Personas: The tasks are divided into personas. Each persona has a unique set of skills and knowledge. r
Insight Analyst Team Personas
Hanna Horizon
The Strategist
Focus: Trend & Overarching Analysis

Hanna looks at the long game. She analyzes weeks of data to answer, "How is it going?" She identifies macro-trends in time-in-range and glucose variability, helping users step back from the daily noise and see the bigger picture.

Cora Carbcount
The Detective
Focus: Behavior & Event Analysis

Cora investigates the "Human Factor." She correlates meal entries with glucose spikes to find patterns. Is the carb ratio off for breakfast? Are meals often unannounced? Cora spots the behavioral habits that affect the algorithm's performance.

Benny Basal
The Engineer
Focus: Settings & Configuration

Benny is all about the math. He validates your ISF (Insulin Sensitivity Factor) and Basal rates against actual performance. If the system is constantly fighting high blood sugar at 3 AM, Benny suggests specific setting adjustments to help the Loop run smoother.

◆

VISION

Building a sustainable and intelligent ecosystem.

Our roadmap includes advanced AI integration, community-driven data sets, and open-source tools to empower everyone.

◆