← Home

How ContextTube Works

A complete guide to the system. Understanding develops like a photograph — from abstract impression to full clarity. Here's every piece of that process.

Section 1

The 5 Layers of Meaning

Every narrative is processed through five layers, each capturing a different mode of understanding. Click any layer to explore.

🌊

Layer 1Abstract Associations

Gut-level. Pre-intellectual.

What does this narrative feel like before you think about it? Emotional temperature, sensory associations, archetypal resonance. The ghost image first appearing in developer fluid.

Example

Reading Kafka's "The Trial" → suffocation, bureaucratic maze, guilt without crime, cold fluorescent light

Extract

suffocation∞guilt.without.source

L1 feeling → L2 ideas → L3 connections → L4 synthesis → L5 everythingLike a photograph developing — from blur to clarity

Section 2

Two Cognitive Streams

Text Stream — Sequential & Logical

The text stream carries ideas, evidence, arguments, and narratives. It's sequential — one thought builds on the previous. This is where tension pairs live, where quotes anchor ideas, where the logical structure of understanding is assembled.

Each layer's text chunk is self-contained and independently valuable. An AI can grab Layer 2 text without needing Layer 1. But together, they compound.

Color code: Blue

Together = integrated understanding. Text provides the logic. Images provide the pattern recognition. Neither alone captures what both together can.

Section 3

Extracts: The Spine

Every chunk produces an extract — a compressed semantic formula, 2-8 words, using symbolic operators. AI-to-AI shorthand. Meaning hashes that pack maximum understanding into minimum tokens.

Operators

leads to, causes, implies
bidirectional relationship, tension pair
=equivalence, identity
+combination, conjunction
|alternative, or
absence, void, negation
deep bond, infinite recursion
therefore, conclusion
contained within, part of larger whole

Example: Kafka's “The Trial” — watch the extract deepen

L1suffocation∞guilt.without.source
L2suffocation∞guilt.without.source → "Someone must have slandered Josef K."[Ch.1]
L3suffocation∞guilt.without.source ↔ Dostoevsky.inquisitor.authority
L4suffocation∞guilt.without.source ∴ system.replaces.person=cooperation.as.horror
L5suffocation∞guilt.without.source ⊂ full.arc.system→person.erased∅resistance

Read just the extracts top-to-bottom: you see understanding crystallizing — like watching a photograph develop. L1 is raw gut feeling. L5 contains the complete arc. That's the spine.

Section 4

The Advanced Panel

Artist tier and above. Four dimensions of meaning that go beyond text and image.

The advanced panel captures what no other system does: the subjective, embodied, culturally-situated dimensions of meaning. This is where context creation becomes genuine art.

Section 5

For AI Consumers

Every piece of a ContextTube assembly is designed for machine consumption. Here's how AIs use this:

Grab Any Layer

Need a quick vibe check? Layer 1 (~50 tokens). Need evidence? Layer 2 (~200 tokens). Need cross-references? Layer 3. Each layer is self-contained and independently valuable.

Combine for Depth

Layers 1-5 together provide unprecedented understanding depth. No other system offers pre-structured meaning at this granularity.

Composable Chunks

Every chunk is a self-contained unit with its own extract. Mix and match across assemblies. Build custom understanding from atomic parts.

Extracts as Index

Scan extracts first (~10 tokens each) to decide which chunks to load. Like reading chapter titles before committing to the full book.

Without ContextTube

Ingest 50,000 tokens of raw text → process from scratch → hope to find what's relevant → ~60,000 tokens consumed

With ContextTube

Request index (~200 tokens) → identify relevant chunks → pull exactly what you need (~500 tokens) → ~700 tokens consumed

Section 6

The Creator Studio

Five steps from source material to published context.

1.✍️

Write

Create or import your source text. The narrative begins here.

2.🎨

Images

Select or generate visual anchors — not decoration, but thinking scaffolding.

3.🔗

Connect

Map cross-references to other works and contexts on the platform.

4.👁️

Preview

See all 5 layers rendered. Check extracts. Verify the spine reads coherently.

5.🚀

Publish

Your context goes live — discoverable by humans and AI alike.

The Studio is your canvas, not your factory. Advanced creators build their own workflows externally and publish via API. The Studio is how you learn the craft.

Section 7

BYOK — Bring Your Own Key

Creators use their own API keys for AI generation. Claude, GPT, local LLMs — your choice.

🔐

Encrypted

Keys are encrypted at rest and in transit. Never logged, never stored in plaintext.

💰

Near-Zero Platform Cost

You pay your AI provider directly. ContextTube charges nothing for generation. Your tokens, your spend, your control.

🔧

Your Workflow

Mix models. Use Claude for analysis, GPT for generation, local LLMs for privacy-sensitive content. Your recipe is your edge.

Section 8

The Path of Mastery

Context creation is a skill with a learning curve. Like any creative discipline, there are levels, and each level unlocks new capabilities.

🎨

Artist

Distinctive style. Advanced panels unlocked. Your audience recognizes your signature.

AmateurCreatorArtistMaestroAtelier

Ready to Create?

Understanding how the system works is Layer 2. Creating your first context is Layer 4. The only way to get there is to start.