Open to projects and collaboration

I build web tools that feel alive — and ship them to production.

Hi, I'm Vitalii Fedin. From real-time computer vision in the browser to interactive simulations and self-hosted AI transcription, every project below is live, working and deployed on a server I set up and run myself.

See the projects Get in touch
Vitalii Fedin
Vitalii Fedin
4live projects, try them now
2.5 yrsof hands-on development experience
20+projects built in total
1self-managed Ubuntu server, HTTPS
Selected work

Four projects. All live.

Different domains, one standard: clean UI, real engineering underneath, and a working demo you can open in one click.

01 — WEB TOOL · SECURITY

FreelancerShield

Protects a client's website preview with removable watermarks until the invoice is paid. Drop the site folder, flip the switches, download a protected copy. Unlock with a code or from your own server.

  • Serverless unlock codes: only a salted SHA-256 hash lives in the page
  • Optional server mode with a Node API to grant and revoke access
  • Live preview, tamper guard, ZIP export, all in the browser
Vanilla JSWeb Crypto / SHA-256MutationObserverNode.jsTauri + Rust
Protect a site →
02 — SIMULATION · EDUCATION

PID Controller Studio

An interactive primer on the most widely used feedback controller in industry. Read the theory, then drag the gains and watch a live plant react: overshoot, oscillation and settling, in real time.

  • Fixed-step physics loop with three process models
  • Live canvas plot of setpoint, process variable and output
  • Zero dependencies, no build step, instant load
Vanilla JSCanvas 2DControl theoryResponsive CSS
Open the simulator →
03 — AI · FULL STACK

File Transcriber

Upload audio or video and get an accurate transcript with timestamps, auto-detection of the spoken language, an optional translation and export to TXT, JSON or PDF. Speech recognition runs on my own server with whisper.cpp, no paid API behind it.

  • Async job queue with live progress and queue position
  • 14 languages with auto-detect, fast or accurate Whisper model
  • Rate limits, daily quotas and auto-deleted uploads
Node.jsExpresswhisper.cppffmpegVitesystemd
Try the transcriber →
04 — COMPUTER VISION · WEBGL

Visual FX Studio

Camera effects that stick to your body in real time. Hands and face are tracked right in the browser, and effects render only inside the regions you choose: finger gaps, palm, eyes, a portal between both hands.

  • 24 body anchors, 8 GPU effects in a single WebGL uber-shader
  • Pinch gestures to drop and drag stationary pins
  • Photo and video capture, works as an OBS source for calls
React 19TypeScriptMediaPipeWebGL2Tailwind 4Vite
Launch the studio →
Services & prices

What I can build for you.

Starting prices. The final price is fixed after a short talk about your project.

Landing page

One page that explains your offer and brings leads

from$300
ready from 7 daysOrder

Interactive web app

Tools, dashboards, calculators, anything users work in

from$900
ready from 14 daysOrder

Telegram bot

Orders, bookings, notifications or automation in Telegram

from$250
ready from 5 daysOrder
Full pricing & cost calculator
How I work

From first message to launch.

  1. 1Message me

    Tell me about the idea through the form or any messenger.

  2. 2Discuss the project

    A few questions in chat or a quick call to understand what you need.

  3. 3Fixed quote

    You get a clear price and deadline before any work starts.

  4. 4Progress updates

    Regular previews, so you see the result while it's being built.

  5. 5Delivery

    The finished project, deployed and handed over to you.

What I bring

From idea to deployed product.

Frontend that performs

React, TypeScript and plain JS. Real-time rendering with WebGL and Canvas, smooth animation loops, and interfaces that stay clean and responsive on any screen.

AI and computer vision

On-device body tracking with MediaPipe and self-hosted speech recognition with Whisper. Turning ML models into features people can actually use.

Backend and DevOps

Node.js and Python: APIs, job queues, child-process workers and Telegram bots. Linux servers, nginx, TLS, systemd services and firewalls, set up and maintained by hand.

Also in my toolbox

I build Telegram bots too.

Private automation tools for myself and clients. They aren't public demos, but I'm happy to show the code.

Remote PC control bot

Controls a Windows PC from Telegram: screenshots, launching installed apps from a paginated menu and running PowerShell commands, locked to a whitelist.

Pythonaiogram 3asyncioMiddleware authInline keyboardsPillow

Music & voice bot

Finds tracks through the YouTube Music API, processes audio with FFmpeg and turns any audio file into a Telegram voice message.

Pythonaiogram 3 routersytmusicapiFFmpeg / pydubSelenium
Under the hood

One server. Everything self-hosted.

No platform magic: every request to this page goes through infrastructure I configured myself on an Oracle Cloud Ubuntu machine.

HTTPS edgenginx with a Let's Encrypt certificate, HTTP redirected to HTTPS:443
Static appsThis landing, PID Studio, FreelancerShield and the Vite build of FX Studio/ · /pid/ · /fx/ · /shield/
Node serviceExpress API under systemd behind a reverse proxy/transcriber/
AI workersffmpeg and whisper.cpp child processes in a FIFO queuewhisper-cli
Contact

Have a project in mind? Let's talk.

Describe the idea in a few words and I'll reply with questions or a quote.