Page cover

Technical Architecture

At the core of Trench Talk is LiveKit, a HIPAA-compliant real-time audio framework that enables encrypted, low-latency voice communication directly in the browser.

Core Components

  • Chrome Extension Injects the Trench Talk UI and detects Solana token pages on supported platforms.

  • Chart Parser Extracts the token’s contract address from Pump.fun, DexTools, Axiom, Photon, and Dexscreener.

  • Room Manager Matches users to the correct voice room based on the token address. Cross-platform syncing ensures users land in the same room regardless of which site they’re on.

  • WebRTC Audio Engine (LiveKit) Real-time encrypted voice communication with sub-100ms latency. No recordings, no storage, ever.

  • Socket Moderation Layer Enables community-based kicking by broadcasting room participants and allowing users to vote out disruptive peers in real time—no central admin required.

  • Username Lock System Assigns a temporary identity locked to the browser for 24 hours to prevent impersonation and maintain speaker consistency.


Privacy & Security

  • End-to-End Encryption (AES-256) All voice traffic is encrypted in transit and never stored.

  • TLS 1.3 Transport Security Ensures encrypted data flows between clients over secure connections.

  • Zero-Knowledge Architecture Voice data is never saved, logged, or recorded. It exists only in the moment.

  • Room Isolation by Contract Each token has its own isolated voice room. No crossover.


Performance Stack

  • Frontend: React 18, TypeScript, Tailwind, Framer Motion

  • Backend: Node.js, Fastify, Redis, WebSocket, JWT

  • Voice Engine: WebRTC via LiveKit with adaptive bitrate and dynamic echo suppression

  • Extension: Built on Chrome Extension Manifest V3

Last updated