N_ NOXGLE.
OPEN PROJECTS · PRACTICAL KNOWLEDGE

Sebastian Wielgosz / NOXGLE

Engineering
in practice.

Open source projects, Linux systems and developer tools — from source code to practical resources.

01 OPEN SOURCE02 LINUX & SYSTEMS03 AI & TOOLS
● ● ●noxgle@hub: ~⌘

➜ ~ cat mission.txt

// build · explore · share

➜ ~ ls ./engineering

open-source/   linux/   automation/
ai-agents/   infrastructure/

➜ ~ ▍

● SYSTEM ONLINENOXGLE / ENGINEERING
01 / OPEN SOURCE

What I build.

All projects ↗
Flagship projectPython / MCP / CLI

ONE ↗

ONE connects local or hosted models to an autonomous Python agent. TUI, headless tasks and RPC offer different ways to work.

ONE — original terminal interface image from the author’s repository
Image from the ONE repository · source ↗

Problem

Working with models, tools and terminal tasks often means switching between separate interfaces.

Use case

Interactive terminal work, task automation and integration with your own tools.

Simplified diagram based on the documentation
  1. TUI / CLI / RPC
  2. AgentSession
  3. Providers + tools / MCP

Source: project README ↗ · Checked 2026-11-16

Local inferenceDocker / CUDA / llama.cpp / Shell

llama ↗

The repository documents a Docker-based llama.cpp deployment with CUDA, model configurations and an OpenAI-compatible API.

Problem

A local model needs a repeatable GPU environment and an interface other tools can use.

Use case

A self-hosted inference endpoint for OpenAI-compatible clients.

Simplified diagram based on the documentation
  1. API client
  2. Docker / llama.cpp
  3. CUDA / GPU + model

Source: project README ↗ · Checked 2026-11-16

MCP integrationPython / MCP / Docker

mcp-web-deepsearch ↗

The server wraps WebSearchAgent: DuckDuckGo, iterative query refinement and content extraction. Available over stdio or streamable HTTP.

Problem

An MCP client needs search that extracts page content rather than returning snippets alone.

Use case

Providing web-search results to MCP-capable clients.

Simplified diagram based on the documentation
  1. MCP client
  2. MCP server / Docker
  3. Search engine → web sources

Source: project README ↗ · Checked 2026-11-16

ENGINEERING / COLLABORATION

How I can help.

Let’s discuss collaboration ↗

From a specific technical problem to a defined scope. Your needs and the work documented in my open projects are the starting point.

01 / ENGINEERING

Linux & automation

Let’s discuss environment configuration, repeatable operations and tasks worth moving into scripts.

Related project: ONE →
02 / ENGINEERING

Local models & infrastructure

Discuss self-hosting approaches, hardware requirements and integration of an inference endpoint with your tools.

Related project: llama →

What should your message include?

Your goal, current environment, constraints and expected scope. Do not send passwords, API keys or confidential logs.

Tell me about your project ↗
02 / KNOWLEDGE

From code to knowledge.

Explore knowledge ↗
03 / VIDEO

Watch and explore.

All videos ↗
2026-10-28

Listing /tmp Folder Contents with ONE

Technical material from the how2 channel. Source description covers: Listing /tmp Folder Contents with ONE. Project relations use only links published by the author.

CLIMCPPython
2026-10-28

Meet ONE

Technical material from the how2 channel. Source description covers: Meet ONE. Project relations use only links published by the author.

CLIMCPPython
LET'S CONNECT

Let’s connect over technology..

Get in touch ↗