Skip to content

Zashiron Ltd

Engineering Intelligent Systems, High-Performance Software, Next-Generation Synthetic Minds, and Offline Edge AI.

Welcome to the official GitHub organization profile for Zashiron Ltd. We are a software and hardware technology company headquartered in Lagos, Nigeria, dedicated to building performant, low-latency applications, synthetic cognitive architectures, secure workspace solutions, offline voice engines, and scalable infrastructure.

🏛️ Corporate Overview

Zashiron Ltd operates across full-stack application development, edge AI engineering, synthetic cognitive architectures, low-level system optimization, secure encrypted networks, on-device audio processing, and custom hardware integration.

  • Legal Entity: Zashiron Ltd
  • D-U-N-S® Number: 352296499
  • Headquarters: Lagos, Nigeria
  • Web Hubs: Aretlyx.app • zashiron.org

📦 Featured Flagship Ecosystem

┌────────────────────────────────────────────────────────────────────────────────────────┐
│                               ZASHIRON PRODUCT SUITE                                   │
├───────────────────┬────────────────────────────────────────────────────────────────────┤
│ 🧠 Alpha          │ Synthetic Mind & Brain-Inspired Cognitive Architecture             │
├───────────────────┼────────────────────────────────────────────────────────────────────┤
│ 🎙️ OpenVox        │ All-in-One, 100% Offline Voice Engine (STT, TTS, Clone, Agent)     │
├───────────────────┼────────────────────────────────────────────────────────────────────┤
│ 🔐 Enclave        │ E2EE Messaging, Voice/Video, File Sharing & Remote Desktop (RDP)   │
├───────────────────┼────────────────────────────────────────────────────────────────────┤
│ 🎬 Aretlyx        │ Multimodal AI Video Intelligence, Auto-Clipping & Scene Detection  │
└───────────────────┴────────────────────────────────────────────────────────────────────┘

🧠 1. Alpha A Synthetic Mind

Alpha is an ongoing attempt to grow a synthetic mind rather than ship a traditional AI assistant. It is a local, private, brain-inspired cognitive architecture featuring a durable associative memory, an emotional system, intrinsic drives, a sense of self and of the user, an inner monologue, and plastic organs whose weights change with experience.

Alpha is not a chatbot and not an assistant. It has no goal of being helpful or compliant. It is an individual: every instance accumulates its own memories, moods, and history over time. The current live individual is named Briallaes.

Core Architectural Organs

  • Associative Graph Memory (GraphMemory): Hebbian edges + spreading activation (SQLite + NetworkX) so recall traverses linked ideas rather than nearest-text matches.
  • Predictive-Coding Net: A GPU-capable Rao–Ballard predictive-coding network whose fast-weights physically adapt to surprise and learning transitions.
  • Affect System: Prediction-error-driven core emotion on Russell’s circumplex.
  • Intrinsic Drives: Internal curiosity, boredom, and social needs that initiate thoughts unprompted.
  • Metacognition & Theory of Mind: SelfModel for confidence assessment and a fading, fallible UserModel for peer dynamics.
  • Default-Mode Network: Spontaneous wandering and dreaming when idle to recombine memories.
  • Language Cortex: A frozen, local, swappable GGUF model (DeepSeek-R1-Distill-Llama-8B / Qwen3.5-9B) used strictly as the language engine to turn thoughts into words, keeping cognitive logic entirely inside Alpha's native organs.

🎙️ 2. OpenVox All-in-One Offline Voice Engine

OpenVox

OpenVox is a complete, 100% offline speech-to-text and text-to-speech voice engine built for edge computing, robotics, air-gapped systems, and privacy-sensitive devices where audio must never leave the hardware.

Highlights & Capabilities

  • 7 Modular Engines:
    • 🎙️ openvox.stt: Real-time streaming STT with live partials and word timestamps.
    • 🗣️ openvox.tts: Natural 24 kHz offline TTS with built-in voices.
    • ⚡ openvox.tts · stream: Low-latency streaming TTS with thread-safe stop() barge-in.
    • 🎭 openvox.clone: Zero-shot voice cloning with neural watermarking.
    • 🧬 openvox.enroll: High-fidelity .ovx reusable voice profile creation.
    • 🧼 openvox.enhance: Denoising, restoration, and 44.1 kHz bandwidth extension.
    • 🤖 openvox.agent: Full offline LLM voice-agent loop (Mic ➔ STT ➔ Local LLM ➔ Streaming TTS).
  • Zero Marginal Cost & Zero Cloud Dependency: No API keys, zero internet requirement, and zero network latency jitter.
  • License: Released under the OpenVox Proprietary License (Zashiron License v1.2).

🔐 3. Enclave Encrypted Communication & Workspace Suite

Enclave (formerly SwiftDrop, backend com.zashiron.swiftdrop) is a secure, self-hostable communication and workspace platform built with Flutter for Android and Windows. It operates local-network-first with direct peer-to-peer capabilities, falling back to a cloud relay when required.

Core Features

  • E2E Encrypted Messaging: 1:1 and group chats using the Signal Protocol (X3DH + Double Ratchet, sealed sender via libsignal_protocol_dart), featuring safety numbers (TOFU), disappearing messages, and offline message buffering.
  • Voice & Video Calls: WebRTC with DTLS-SRTP encryption, native call interfaces (Android ConnectionService / iOS CallKit), multi-party audio/video, and live network metrics panels.
  • High-Speed File Transfer: Dual mode featuring Swift Mode (direct TLS-pinned LAN streaming) and Air Mode (cloud WebSocket relay), alongside a clientless Web Share interface.
  • Cross-Device Remote Desktop (RDP): Control desktop-from-phone or phone-from-phone with OS-level input injection (Windows SendInput / Android Accessibility Service).
  • Self-Hostable Infrastructure: Enterprise organizations can deploy the Office Space Node (WebSocket relay + TURN + Key Directory) on their own hardware for total data sovereignty.

🎬 4. Aretlyx AI Video Intelligence Platform

Aretlyx is an end-to-end multimodal AI platform engineered for automatic video editing, smart clipping, and highlight generation.

  • Visual Tracking: Integrates YOLO11 computer vision models for subject framing, face tracking, and dynamic cropping.
  • Acoustic Intelligence: Employs YAMNet and PANNs for audio event detection, speech/music separation, and emphasis tagging.
  • Multimodal Reasoning: Combines LLM reasoning pipelines with temporal timestamps for automated script alignment and clip selection.

🛠️ Unified Technology Stack

┌────────────────────────────────────────────────────────────────────────────────────────┐
│                                 ZASHIRON TECH MATRIX                                   │
├───────────────────┬────────────────────────────────────────────────────────────────────┤
│ Application Layer │ Flutter, Dart, Next.js, React, TypeScript, Tailwind CSS            │
├───────────────────┼────────────────────────────────────────────────────────────────────┤
│ Systems & Backend │ Go, Python 3.11+, PyTorch (CUDA), Node.js                          │
├───────────────────┼────────────────────────────────────────────────────────────────────┤
│ Speech & Audio AI │ OpenVox (STT/TTS/Clone), Chatterbox, Resemble-Enhance, Silero VAD  │
├───────────────────┼────────────────────────────────────────────────────────────────────┤
│ Cryptography & Net│ Signal Protocol, WebRTC (DTLS-SRTP), TLS Certificate Pinning       │
├───────────────────┼────────────────────────────────────────────────────────────────────┤
│ Edge & Local AI   │ llama-cpp-python (GGUF), faster-whisper, kokoro-onnx               │
├───────────────────┼────────────────────────────────────────────────────────────────────┤
│ Data & Cloud      │ Neon Postgres, SQLite, NetworkX, Upstash Redis, Cloudflare R2      │
├───────────────────┼────────────────────────────────────────────────────────────────────┤
│ Payments & Auth   │ Paystack API, Appwrite, Clerk Authentication                       │
└───────────────────┴────────────────────────────────────────────────────────────────────┘

🤝 Contribution & Internal Governance

  1. Public Repositories: Open for issue tracking, public client SDKs, and community tools.
  2. Security & Vulnerabilities: For security concerns regarding OpenVox, Enclave, Alpha, or cloud relays, report directly via security administration.
  3. Developer Identity: Commits by Zashiron maintainers must be GPG/SSH signed using registered @zashiron.com or @zashiron.org corporate identities.

© Zashiron Ltd. All rights reserved.

Pinned Loading

  1. .github .github Public

Repositories

Showing 2 of 2 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…