Offline Content Management System for Remote-Viewing Sessions
Heichalot-CMS is a lightweight, offline-first system designed to collect, structure, and transform remote-viewing session data into reusable, distributable content.
The core idea is simple:
Take raw inputs (transcripts, sessions, videos) → convert them into a structured narrative format offline. That can be on the computer/laptop/phone, on an SD-Card, USB-Drive or a Storage. Backed up in case they are ever removed from the Internet.
🧩 The story.md concept
Each session is stored as a single markdown file:
---
kind: story
source: youtube
source_video_id: XXXXX
---
# Session Title
""" NARRATOR
Session content appears here as a clean, readable narrative block.
"""
This format is:
- human-readable
- AI-readable
- easy to render into video, PDF, HTML (including 3D world ThreeJS)
- stable across tools and environments
📁 Core CMS Structure
cms/
entry-0000001/
story.md
assets/
entry-0000002/
story.md
assets/
Each entry represents one session.
- story.md → primary narrative
- assets/ → supporting data (original transcripts, translations, metadata)
This allows the system to scale cleanly while keeping each session self-contained.
🛠 Tools Layer
tools/
ytv2cms.py
renderpdf.py
renderhtml.py
startproduction.py
extensions/
These scripts operate on the CMS data and convert it into different forms.
🔌 Extensions Layer
tools/extensions/
registry.py
translation_google.py
translation_libre.py
translation_argos.py
ytdlp_metadata.py
diarization.py
Extensions provide optional capabilities:
- translation - translate from source language to local language, ie German->English, English->German
- metadata extraction
They are loaded dynamically and do not affect the core system if absent.
⚙️ Key Scripts
👉 ytv2cms.py
Converts a YouTube video into a structured story.md.
- extracts transcript
- formats into narrative
- optionally translates
- writes clean CMS entries
This bridges raw video → structured dataset
👉harvestcreatorcontent.py
Scans an entire creator/channel and builds a corpus. Such as Future Forecasting Group, Farsight, Bulgarian Remote-Viewers, Edward O'Riordan, Down the Rabbit Hole, David Powell.
- collects all video URLs
- builds a producer manifest
- feeds videos into the CMS pipeline
This enables bulk ingestion at scale, not just single sessions
👉 renderpdf.py
- Turns story.md into printable PDF reports.
- Converts session data into distributable documents
👉renderhtml.py
Renders sessions as web-ready HTML.
- Enables browser-based viewing and embedding
👉startproduction.py
- Prepares a production workspace from a CMS entry.
- Connects structured data to video/visual pipelines
registry.py (Extensions)
Dynamic capability loader.
- detects available extensions
- enables features without breaking core tools
👉 This allows free vs premium capability layers without changing the main system
Translation Extensions
- Google (high-quality, paid)
- Libre / Argos (free, offline-capable)
Enables multilingual session archives
👉diarization.py (future)
Speaker separation (interviewer vs viewer)
- Moves from transcripts → structured dialogue
🧭 What’s new here
This system introduces a different approach:
- sessions are treated as structured data, not just media everything is offline-capable
content is portable and reproducible.
- ingestion, transformation, and rendering are separated
The result is a pipeline that can:
- collect remote-viewing sessions at scale
- normalize them into a consistent format
- prepare them for analysis, publishing, or further processing
🔚 Summary
Heichalot-CMS is not just a storage system—it is a content transformation pipeline.
It turns scattered session recordings → structured, reusable knowledge
If there’s interest, I can share examples of generated entries or demonstrate how a full channel is converted into a CMS dataset.