Circadian entrainment, audio journaling, and AI as a cognitive prosthetic.
A prologue project exploring Sesefus — AI-augmented recovery tools for individuals in therapeutic settings and beyond.
View the site: https://zychs.github.io
This repository hosts the project prologue and documentation for Sesefus, a comprehensive system built around:
- Circadian entrainment through structured audio journaling
- Rich alarm scheduling with profile-based management
- Local-first AI inference for cognitive augmentation
- Voice-first interfaces for accessibility in restricted environments
index.html— Full prologue essay with design and typographic stylingsupport.js— Minimal supporting scriptREADME.md— This file
To run locally:
# Clone the repository
git clone https://github.com/zychs/zychs.github.io
cd zychs.github.io
# Serve with a local server (Python)
python -m http.server 8000
# Or with Node.js
npx http-serverThen visit http://localhost:8000 in your browser.
- Host + Client Model — Scalable audio capture and processing
- Local Inference — vLLM integration for local language model support
- Vault Storage — Encrypted audio and artifact databases
- TCP-based Communication — Low-latency client-server messaging
- Language: HTML5 + CSS3, JavaScript
- Fonts: Outfit (body), JetBrains Mono (code/headings)
- Styling: Custom gradient backgrounds with responsive typography
- Core Engine: Zig-based Circadia/Sesefus alarm system
- Inference Layer: Crow-9B-HERETIC via local vLLM
- Infrastructure: LeadLogic-Engine for embedding and reflection
For Everyone · Voice-First · Unhurried · Negentropic