Art2Dec SoftLab

CI/CD, SaaS, PaaS, IAC and OSS Solutions

Our non-profit softlab is committed to promoting solutions of various complexity levels based on cloud and oss technologies. The basis of our technology stack are solutions based on own SaaS framework, k8s solutions (AWS/GCP/on-premise), docker farms and components related to CI/CD (Continous Integration/Continous Development). Two factors are fundamental in the implementation of our decisions. The first factor is the desire to use in our solutions the widest possible range of open technologies to optimize the price-performance ratio. The second factor is our desire to implement turnkey integration solutions with the entire range of products necessary for the implementation and use of integrated and hybrid systems. The latter include development tools, control tools, debugging tools, distributed RDBMS/NoSQL/graph database systems, grid systems, various middleware technologies, protection tools, performance monitoring tools, and much more.

Blog Slides Forum Live Demos

Latest Developments & Open-Source Projects

Art2Dec SoftLab R&D has produced a complete AI ecosystem for Linux: a new shell, workflow engine, GUI tools, network scanner, AI agents, and scientific applications. All projects are open-source and actively maintained.

mshell Ecosystem — Unix Shell for AI & Mathematics

mshell is a new Linux/Raspberry Pi shell written in C from scratch, purpose-built for AI workflows and mathematical computing. Integrates directly with local LLMs (Ollama, phi, llama, mistral) and cloud LLMs (Claude, ChatGPT). Automatically validates and executes AI-generated code. Supports Lua, Python, Bash, Go, Rust, C, C++ scripting.

The mshell Ecosystem includes: mshell Shell · mshell Studio (GTK GUI editor) · Library Manager · Knowledge Base & Validator · MEL Editor · mshell Workflow engine.

Documentation & Guides:

mshell — new Linux shell for AI and mathematics (Mar 2025)
mshell Ecosystem — short description of main components
mshell v1.4.1 Cheatsheet
mshell Functions & Structures — Complete QA Analysis
FOR / WHILE / IF Guide for mshell
mshell — Inter-Language Execution Reference Guide
mshell — Working with PDF files (new option)
Simple example: using mshell Ecosystem
Two LLM Pipeline experiment

mshell Workflow — Multi-Language Agentic Pipeline Engine

Programmable pipeline environment that orchestrates multi-language code execution through a visual flow diagram interface. Complex applications are built as directed acyclic graphs of code blocks — outputs of one block become named inputs of the next. Supports all five canonical LLM agentic workflow patterns.

24 workflow patterns documented across 9 languages: mshell, Bash, Python, C, C++, Rust, Go, Lua, and Unified. Polyglot benchmark comparing performance across all languages included.

Real-world applications built with mshell Workflow: Nettopo3 network scanner, AI Email Agent, matrix multiplication benchmarks, generative & materialized documents, and more.

Guides & Reference:

mshell Workflow Guide (Feb 2026)
mshell Ecosystem — Tools & Components Reference for Workflow
Workflow Patterns Reference Guide — Part I (P1–P12)
Workflow Patterns Reference Guide — Part II (P13–P24)
Polyglot Benchmark Report (Mar 2026)
NxN Matrix Multiplication Performance Study
AI Agentic System Analysis: mshell Workflow Q&A (Apr 2026)
mshell Workflow — Generative & Materialized Documents

Language-specific Pattern Guides (P1–P24):

Unified (all languages) — Complete Reference
Pure mshell language Patterns · Bash · Python · C · C++ · Rust · Go · Lua

Nettopo3 — circular network topology graph

Nettopo3 — network scan with host detail panels

Nettopo3 — Real-Time Network Topology Scanner

Standalone network discovery and visualization tool built with Go, C/SDL2, and Bash via mshell Workflow. Continuously scans IP ranges, discovers active hosts, and renders a live interactive graph with latency color-coding. Supports 243+ hosts across multiple subnets simultaneously.

OS detection — Linux, Windows, Network Device, macOS · Wi-Fi detection by ping jitter · Real-time diagnostics — ping, nmap, traceroute, ARP, netstat in simultaneous windows · Zero-configuration — two binaries and a shell script, no database or agents required.

Built using the mshell Workflow engine — Go for concurrent scanning, C/SDL2 for 60fps hardware-accelerated rendering, Bash for orchestration. From launch to full /24 subnet picture: under 15 seconds.

Nettopo3 — User Guide v3.0

mshell Workflow — Five Canonical Agentic Patterns

mshell Workflow implements all five canonical LLM agentic workflow patterns used in modern AI system design: Sequential, Parallel, Conditional, Loop, and Hierarchical. Each pattern is fully documented with working examples across all supported languages.

These patterns form the foundation for building complex AI agents and pipelines — from simple sequential processing to multi-agent orchestration systems.

Implementation of all five canonical agentic workflow patterns in mshell
Workflow Patterns in LLM Agentic Systems

AI Intelligent Email Agent for IMAP/POP3 Mail Systems v1.0

Complete turnkey AI-powered email agent written in Go, C, and Bash using the mshell Workflow engine. Connects to any IMAP/POP3 mail server, intelligently processes incoming messages using LLM reasoning, and automates responses based on configurable rules.

Supports local LLMs (Ollama, phi, llama, mistral) and cloud LLMs (Claude, OpenAI). Includes client interface, monitoring dashboard, and manual mode. Full technical documentation and user guide included.

AI Intelligent Email Agent — Technical Documentation & User Guide

Everyday Programmer Notebook on Raspberry Pi

MEL Editor & Every Day Programmer's Notebook

MEL Editor — terminal-based text editor written in pure C, no external dependencies. Runs on Linux, macOS, Raspberry Pi (ARM32/64). Syntax highlighting for 15+ languages. Designed for Kubernetes pods, Docker containers, IoT devices.

Every Day Programmer's Notebook — cross-platform GUI markdown editor and notebook, part of the mshell Ecosystem. Runs standalone on macOS. Features inline markdown rendering, emoji support, multiple export formats, code snippet management.

Documentation:

MEL Editor — User Guide
Every Day Programmer's Notebook — Main Features (Jan 2026)
Notebook — Format & Styling Support
Notebook — Complete Export Formats Report
8 bytes (64 items) Bash Snippets Collection
8 bytes (64 items) Python Snippets Collection

Linux LLM Evaluation Framework for Local Use

Comprehensive framework for evaluation and integration of large language models in a local on-premise environment. Comprises six key modules:

1. One-Pass Compilation Module — integrates the Art2Dec All-in-One compiler supporting Go, Java, C++, Python. Includes Prompts Repo, Evaluator, Analyzer, and API module.

2. Data Ingestion Module — handles plain and binary files, databases, and programming channels for structured data preprocessing.

3. Ollama Module — central hub for managing LLM interactions, connecting with the LLM repository and coordinating APIs.

4. LLM Repository — structured storage for different versions and types of LLMs for rapid testing and deployment.

5. Chat & CMD Chat Modules — interactive interfaces for LLM sessions with command-line string array manipulations and history management.

6. APIs & Integrations — prompts, evaluation, analysis, and Ollama API integration ensuring seamless communication between all components.

Linux LLM Evaluation Framework — Full Description

LLM Evaluation Framework Architecture Scheme

AI Applications & Scientific Publications

Constraint-Driven Semantic Control — research on LLM behavior under progressive lexical restrictions using mshell Workflow. Experiments in AI safety and controllability.

Art2Dec All-in-One Compiler — unified compiler supporting Python, Go, Java, C++, C in a single environment. Integrated with mshell Ecosystem.

Scientific applications: Quantum Beating in Hydrogen Atom · Machine Learning Taxonomy Diagram · Scientific publications with mshell · Platonic Solids construction verification · Data visualizations with LLMs · 7D Pi-Storage System · Local web server testing with mshell.

AI Applications & Research:

Constraint-Driven Semantic Control (Apr 2026)
7D Pi-Storage System
Visualizations with mshell and LLMs (Claude, ChatGPT, Ollama)
Webscraping with mshell and LLM
Declarative Diagnostics & Monitoring with local LLMs
Launching local web servers for rapid testing using mshell

Scientific Publications:

Predictive Modeling of Quantum Beating in Hydrogen Atom
Machine Learning Taxonomy Diagram
How to create scientific publications with mshell v1.4.1
How to correctly construct all five Platonic solids
Using mshell and local LLMs to create non-scientific publications
Urban Balcony Garden Design Project — example with AI

Technical Reference Cards & Mind Maps

Original visual reference cards and mind maps covering OS command mappings, programming languages, databases, APIs, and more. Created as quick-reference cheat sheets for engineers working across multiple platforms and technologies.

Collection includes: OS command maps (Ubuntu, Windows CMD, PowerShell, macOS Sonoma) · Operating Systems mind map · Programming Languages mind map · Databases mind map · REST API operations · RESTless API operations.

Part of the Art2Dec SoftLab Reference Cards collection — the most complete collection of technical reference cards available at art2dec.co/refcards.

Operating Systems

Programming Languages

Databases

REST API Operations

RESTless API Operations

Full project portfolio, documentation, and live demos:

Blog & Documentation GitHub LinkedIn

Below you can find links to the main technologies that we use. Conceptually, they are divided into ready-to-use Applications and integrated DevTools, Ecommerce solutions, Platforms, SemanticsWeb, Based Technologies, Web2.0 tools. We use them, but are not limited by them and can always consider the possibilities of other products at the request of the customer. Technology is constantly changing and improving. The basis for moving forward is open technologies, since it is the basis and foundation of all proprietary solutions at its core.

CI/CD Solutions

To implement the diferent pipelines we use the following products. Learn more...

SaaS Solutions

To know how we build SaaS solutions to use at cloud or on-premise. Learn more...

PaaS Solutions

We use different PaaS and their components based on AWS Services, GCP and Azure. Learn more...

OSS Solutions

All our solutions are based on open source products. Learn more...

Advertising: We have implemented and are promoting the AYNR Internet channel information collection platform technology based on the means of collecting information, its indexing and visualization. How it works - you can look at demonstrations ( sample1, sample2, sample3, sample4, sample5 ) We are looking for investors who want to participate in supporting the development of technology and implementations for different customers.

“The AYNR Channel Information Collection Platform technology is the best choice for automatically maintaining your blog using your simple relevant search rules.” Nickolas M. Developer

“Using the AYNR platform, we have implemented several online channels that help us view the latest news in our area of expertise.
The best thing is that we can include, edit various information and its blocks themselves." Leonard S. Marketer

"I was able to organize for myself a collection of songs of my favorite author without any problems. .Also, I made a great selection of jokes, divided into categories." Mike B. Teenager

Make your own blog in a few clicks! AYNR platform helps you put right rules to automate search an information at open sources. Here is an example of AYNR platform realization for the collection of relevant information and an output in the required format, with indexed search and info collection: The subject of relevant search is an information about Great German Philosopher Immanuel Kant on youtube channels. Also, you can use other sources for relevant search.

Make your own blog in a few clicks! AYNR platform helps you to get and automate your noble relevance of information, implement structured output of information to your Internet channel. Here is an example of AYNR platform realization for the collection of relevant information and its output in the required format, with indexed search and info collection: The subject of relevant search is an information about Ancient Greece on youtube channels. Also you can use other sources for relevant search.

Make your own blog in a few clicks! Here is an example of AYNR platform realization for the collection of relevant information and its output in the required format, with indexed search and info collection: The subject of relevant search is an information about Ancient Rome on youtube channels. Also you can use other sources for relevant search.

Make your own blog in a few clicks! Here is an example of AYNR platform realization cut of the collection of relevant information and its output in the required format, with indexed search and info collection: The subject of relevant search is an information about Russian Bard Oleg Mityaev on youtube channels. Also you can use other sources for relevant search.

Make your search fast! AYNR platform helps you to make indexing. Make your search easier and more convenient, according to the information relevance interesting you. Collect information, the system will index it, use a quick relevant search, if necessary, re-index it. Put the sources to search an information based on open sources to get your info based on relevance of information, implement structured output of information to your Internet channel. Here is an example of platform searching engine component to use.

Most full collection at Internet:

REFERENCE CARDS

FREE TO USE

AYNR is one of the easiest releavance search development tools platform available today. It also gives a freedom to develop as many websites and Internet Channels as you like given the fact that it is an OSS platform completely.

MOBILE FRIENDLY - no special actions required, all sites you make with AYNR are mobile-friendly. Output streams of information adapted to their use with mobile devices. Your internet channel will be very use responsive for you and others.You don't have to create a special mobile version of your site, it will adapt automagically. Try it now!
EASY AND SIMPLE - cut down the development time with drag-and-drop website builder. Drop the blocks into the page, edit content inline and publish - no technical skills required. Simple rules for relevant search. Try the result!
UNIQUE STYLES - choose from the large selection of latest pre-made ready themes, blocks and plugins to realize the internet channel - full support of latest technologies to use, full-screen intro, bootstrap carousel, content slider, responsive image gallery with lightbox, parallax scrolling, video backgrounds, hamburger menu, sticky header and more. Try the result!

Addresses

1037 Howell Pl,
Aurora, IL 60505

Contacts

Email for support: igor.lukyanov@appservgrid.com
Phone: +1 (510) 9904724
Fax: +1 (510) 9904725

Links

Art2Dec Gallery
Art2Dec SoftLab Slides System
Art2Dec Visualisation Desk
Art2Dec Phorum
Art2Dec Image Collection
Art2Dec Own Cloud
Art2Dec Refcards Collection
Art2Dec GitHub
About Author

Feedback

Please send us bug reports, suggestions!
igor.lukyanov@appservgrid.com