ENΧ’Χ‘Χ¨Χ™ΧͺFranΓ§ais

Station V β€” The Toolkit

Software I
Actually Recommend

Not a list of everything that exists. These are the tools that survived contact with real projects β€” the ones I'd reinstall first on a fresh machine.

Engineering

Dev & Build

The daily stack β€” where the code gets written, containerised, and shipped.

🐳

Docker

Container runtime for services, databases, and local model hosting.

WhyThe cheapest way to run something complicated without letting it move into your operating system permanently.

πŸ¦€

Tauri

Desktop app shell using the OS webview instead of a bundled browser.

WhyNative-size binaries with a web front end. Electron's ergonomics without Electron's install footprint.

πŸ”·

.NET Core & C#

The backend language and runtime behind most of my enterprise work.

WhyStrong typing, genuinely good tooling, and cross-platform since Core. Boring in the way infrastructure should be.

πŸ…°οΈ

Angular

Opinionated front-end framework for structured, long-lived applications.

WhyThe opinions are the feature. On a codebase several people maintain for years, convention beats flexibility.

πŸ€–

Kotlin & Android Studio

Mobile development, including large-scale Android deployment work.

WhyKotlin removed most of Java's ceremony without abandoning the ecosystem underneath it.

πŸ”

Elasticsearch & Kibana

Search and log analytics over large document and event volumes.

WhyWhen the dataset outgrows SELECT ... LIKE, this is the honest next step rather than a bigger database.

Local AI

Models That Run On My Machine

Running models locally is not about saving API costs. It is about working on material that should never leave the machine it lives on.

🧠

Gemma (self-hosted)

Open-weight LLM served locally inside a Docker container.

WhyProprietary workflows where sending the input to a third party is simply not an option. The model comes to the data.

⌨️

Claude Code

Agentic coding in the terminal β€” reads the repo, edits files, runs the tests.

WhyAutocomplete guesses the next line. This one holds the whole project in view, which is a different category of useful.

Workflow

Knowledge & Shipping

The unglamorous layer: where notes live, and how a change gets from my laptop to the open internet.

πŸ—‚οΈ

Obsidian

Local-first knowledge base β€” plain Markdown files, linked together.

WhyThe notes are just files on disk. Whatever happens to the app, the decade of writing stays readable.

☁️

Cloudflare Pages

Static hosting with serverless functions, deploying straight from a Git push.

WhyThis site runs on it. Push to master, live in thirty seconds, no build pipeline to maintain.

πŸ™

Git & GitHub

Version control, and the deploy trigger for everything published here.

WhyThe commit history is the only project documentation that never goes stale.

Built With This Stack

Featured Projects

Where the tools above stop being a list and start being a product.

πŸ”Ž AppSearch

High-performance local file indexing for macOS. Byte-level matching and a first-byte prefix index keep search responsive across very large file trees, with incremental updates driven by filesystem events rather than rescans.

macOSFSEventsIndexing

🧠 Local AI Workflows

Locally hosted LLMs (Gemma) embedded into proprietary workflows using Docker β€” inference on sensitive material without any of it leaving the machine.

GemmaDockerSelf-hosted

🎡 yossinew.com

This site. Hand-written HTML, CSS, and JavaScript β€” no framework, no build step, no npm. Serverless functions handle the parts that genuinely need a server.

Vanilla JSCloudflare PagesNo build step

Toolbelt

C#.NET CoreSQL Server AngularKotlinAndroid Studio TauriDockerElasticsearch KibanaGitCloudflare