Gaming Voice / SaaS / Real-time

Zero Lag. Full Squad.

A lightweight voice communication platform for gaming squads that eliminates the bloat of Discord and the unreliability of native platform voice chat. Session-based, no installation, instant voice coordination.

Astro React LiveKit WebRTC Voice Processing Gaming Session-Based Real-time
Project Squadar
My Role Product Designer & Frontend Developer
Year 2024 — Present
Scope UX, UI, Frontend, Voice Engineering

Built for Speed

<2s
Connection Time

From entering a code to speaking with your squad. No accounts, no downloads, no server browsing.

0
Installation Required

Pure web application. Runs in any modern browser on desktop and mobile. Zero friction for guest users.

100%
Voice Control

ClearMic processing, 3-band EQ, push-to-talk, and Voicemod integration built into the client.

01 / The Problem

Voice Chat Bloat

Discord consumes excessive RAM and packs features that nobody uses during a match. Native platform voice chat (Steam, consoles) fails frequently and offers zero control. Squads waste minutes configuring when they should be playing.

02 / The Solution

Session-Based Voice

A pure web app built with Astro and React islands. Enter a room code, get a LiveKit token, and you're speaking in under two seconds. No accounts, no server lists, no persistent bloat. Sub-channels let you split the squad without losing the main room.

03 / The Result

Instant Coordination

Voice communication that feels like a feature of the game, not a separate app. Zero setup, full voice control, and a dark space UI that doesn't fight for attention during gameplay.

The Challenge

Voice chat that doesn't get in the way

Every voice chat tool on the market was designed for social hanging out, not competitive coordination. Discord wants you to build a community. Steam wants you to stay in their ecosystem. Neither prioritizes the thing that actually matters in a squad: getting everyone talking, fast, with clear audio and zero distraction.

The challenge wasn't building another chat app — it was stripping voice communication to its absolute essentials and making it feel instantaneous.

Comparative Analysis

Discord Startup 2-5 min (server browsing)
Squadar Session Start <2 seconds

Discord RAM Usage 300MB — 1GB+
Squadar RAM Usage Browser tab only

Voice Processing None (or paywalled)
Squadar Voice Control ClearMic + EQ + PTT built-in

Technical Architecture

Frontend Framework Astro + React Islands
Real-time Voice LiveKit (WebRTC SFU)
Voice Processing Web Audio API
Voice Modification Voicemod V3 API
Typography Rajdhani + Inter + JetBrains Mono
Design Tokens Dark Space + Cyan + Magenta
Product Implementation

Session-based voice. No accounts, no friction.

Squadar eliminates the concept of "servers" entirely. Instead, it uses temporary sessions: one person creates a room code, shares it with the squad, and everyone joins directly. When the session ends, the room ceases to exist.

  • 1
    Enter a nickname and room code. That's it. No sign-up, no email, no OAuth.
  • 2
    Get a LiveKit token, connect to the SFU, and you're speaking in under two seconds.
  • 3
    Need a private channel? Create a sub-room within the session without losing the main room.

Voice Engineering

Full Control Over Your Voice

ClearMic Processing

Real-time audio enhancement that cleans up mic input before it reaches the squad. Toggle on or off with one tap.

Built on Web Audio API with a 3-band EQ (200Hz bass, 2.5kHz mid, 5kHz treble) that lets you tune your voice for maximum clarity without external tools.

Push-to-Talk

Toggle PTT mode with a single click. Hold space bar to speak, release to mute. No accidental noise bleeding into the squad's comms.

Designed for competitive environments where open mics are a liability. The space bar binding is intuitive and doesn't require tabbing out.

Voicemod Integration

Full Voicemod V3 API integration. Detect when the desktop app is running, browse available voice presets, and switch voices remotely.

No need to alt-tab to change your voice. The status indicator shows when Voicemod is connected, checking, or offline.

Design System

Dark Space + Cyan + Magenta

Palette

Base

Surface

Elevated

Overlay

Cyan

Magenta

Success

Error

Typography & UI

Rajdhani

Display, headings, numbers

Inter

Body, buttons, labels

JetBrains Mono

Channels, metadata, technical states

Key UI patterns: speaking indicators with cyan glow, online badges with green dot, sub-channel chips with hash icon, and voice panel with real-time EQ visualization.

Product Leadership

From Concept to Working Product

Architecture & Engineering

Designed and built the entire frontend in Astro with React islands for interactive components. LiveKit integration for WebRTC voice with custom token management and sub-channel routing.

Voice Processing

Built the audio pipeline: Web Audio API for ClearMic processing and 3-band EQ, Voicemod V3 API integration for remote voice preset control, and PTT mode with keyboard binding.

Design System

Created the complete dark space visual language: palette, typography (Rajdhani/Inter/JetBrains Mono), component patterns, and interaction states for all voice UI elements.

"This project demonstrates my ability to connect design, technology, and business."

Looking for a similar solution?

If you're looking for a hybrid professional who can design and build functional digital products, I'm ready to collaborate.