CoreChat

High-Performance P2P Moderated Messaging

Serverless, encrypted, and moderated by CensorCore.

Overview

CoreChat is a production-ready communication platform that utilizes WebRTC to establish direct connections between users. Unlike traditional chat apps, CoreChat does not store your messages on a server. Everything happens directly in your browser.

Key Features

View the project on GitHub:
GitHub Logo CoreChat Repository

Technical Overview

CoreChat operates as a Static Web Application (SWA) with a decentralized signaling model.

The Connection Lifecycle

  1. Signaling: You receive a unique 6-character ID. When you share this with a peer, a "handshake" occurs via a signaling server to exchange session descriptions.
  2. NAT Traversal: STUN (Session Traversal Utilities for NAT) servers identify your public IP address to navigate through home firewalls.
  3. DataChannel: Once the handshake is complete, a direct encrypted tunnel is opened. From this point on, the signaling server is no longer involved in the conversation.

Safety Standards

To prevent abuse in a serverless environment, CoreChat implements advanced client-side safeguards:

Summary Architecture Moderation Security DevOps Requirements

CoreChat Production Documentation

Executive Summary

CoreChat is a high-performance, serverless peer-to-peer messaging framework designed for secure, private, and moderated communication. By utilizing the WebRTC protocol, CoreChat eliminates the need for intermediary servers to store or relay message data, ensuring that all conversations remain strictly between the participating endpoints.

Project Vision and Architecture

The primary objective of CoreChat is to provide a zero-footprint messaging solution where privacy is maintained by the architectural design. The application functions as a Static Web Application, executing all logic locally in the browser.

Network Topology and Signaling

  1. Signaling: The client connects to a global signaling server to register its unique 6-character ID.
  2. NAT Traversal: The system utilizes STUN servers to identify public IP addresses and navigate firewalls.
  3. Peer-to-Peer Tunneling: Once the handshake is complete, data flows directly from one browser's memory to the other.

CensorCore Moderation Engine

A critical component is the integration of the CensorCore Library, acting as a local firewall for text content.

Content Filtering Logic

  • The string is passed to the CensorCore validator before transmission.
  • The engine checks content against a weighted dictionary of prohibited terms.
  • Violations are intercepted and discarded before reaching the network layer.

Fuzzy Match Reporting

  • Calculates Levenshtein distance between flagged text and prohibited wordlists.
  • Allows the software to catch bypassed filters (leetspeak) dynamically.

Security and Encryption Standards

  • DTLS: All data is encrypted using Datagram Transport Layer Security.
  • Perfect Forward Secrecy: Keys are negotiated for each session.
  • Local Privacy: Message history is volatile; refreshing the browser wipes all data from RAM.

Advanced Feature Specifications

Progressive Suspension System

An automated anti-abuse system tracks user behavior: Initial violations result in local warnings, while repeated violations lead to timed suspensions (1–3 minutes) managed via browser timestamps.

Developer Operations

A hidden administrative portal is embedded for overriding suspensions. Users can trigger the authorization prompt using the keyboard sequence Alt + Q + W.

Technical Requirements

  • Browser: Modern evergreen browsers (Chrome 60+, Firefox 55+, Safari 11+, Edge 79+).
  • Connectivity: Support for WebRTC protocols.
  • Protocol: Must be served over HTTPS for peer discovery permissions.

Version History

v1.0.0 Production Release

CoreChat is a high-performance, serverless peer-to-peer messaging platform designed to provide secure and moderated communication without the need for a central database. By leveraging WebRTC for direct browser-to-browser connections and the CensorCore engine for real-time content filtering, it ensures that all conversations remain private, encrypted, and safe from abuse.

  • Stable WebRTC DataChannel implementation via PeerJS.
  • Full integration with CensorCore Moderation Engine.
  • Automated Progressive Suspension anti-abuse system.
  • Fuzzy-logic detection for "leetspeak" filter bypassing.
  • Zero-footprint SWA architecture (no accounts or logs).
View Release on GitHub

About the Developer

Derrick Richard is a student developer from Poth, Texas (Class of 2029) dedicated to building privacy-focused tools and lightweight libraries.

Other Projects

Visit Official Portfolio