Create an image, Using my attached documents that talk in-depth about my app Chunk, how it works, its features and the Design System. I want

Image-generation prompt: Chunk — The Personal Knowledge Engine

Create a polished 16:9 editorial product-diagram illustration for Chunk, titled:

Everything you learn, finally connected.

Your Personal Knowledge Engine

Use Chunk’s “Paper & Ember” visual system from [[Chunk Design System v2 — “Paper & Ember”]]: warm cream paper background, tactile index-card surfaces, subtle paper grain, warm brown ink, and a rare ember-orange accent. The composition should feel like a sunlit study desk—not a dark technical dashboard.

Central visual metaphor

At the center, show a large, softly three-dimensional rounded cube / squircle “Chunk” — the core Personal Knowledge Engine. It should look like a warm cream paper object with a thin ink outline, gentle warm shadow, and a small ember-orange internal glow. Around it, small rounded “chunks” travel inward, connect, and become a living graph.

The image should clearly tell this story:

Capture → Create → Index & Connect → Retrieve → Compound

Layout

1. Left side: “Everything flows in”

Arrange a scattered but organized set of source cards/chunks flowing from the left toward the central Chunk engine, joined by thin animated-looking ember connection lines.

Include these labeled sources, each as small tactile paper cards with simple restrained icons:

• Share Sheet — phone / Safari article

• Web Clipper — browser page

• Email to Chunk — envelope

• Paste / Drop — URL, text, image, PDF

• Documents — PDF, DOCX, EPUB, spreadsheet

• Notion — connected workspace

• Voice / Audio — waveform or microphone

Use labels such as:

• “Save anything”
• “Capture from anywhere”
• “Articles · PDFs · links · images · ideas”

Make these objects feel like loose study materials arriving at the desk.

2. Top and lower sides: “You and Chunk create knowledge”

Surround the central engine with a ring of Chunk feature cards, each visually distinct but cohesive. These are not merely inputs: they create durable knowledge objects that flow into the engine.

Feature cards:

• Chat
  Prompt bubble → cited answer → “Save as note”
  Include a small “Used in answer” connection cue.

• Notes
  A warm note card with visible wiki syntax: [[Connected ideas]]

• Deep Research
  Several source cards funnel into a structured, cited report.

• Artifacts
  One source chunk branches into:

  • Notes
  • Summary
  • Flashcards
  • Quiz
  • Concept Map

• Collections
  A shelf/grid of selected source cards forming a focused project workspace.

• Automations
  A small recurring loop / clock glyph feeding a cited digest or research update into the system.

• Memory
  A small, calm profile-memory card: “What matters to you, carried forward.”

These feature cards should visually “deposit” their outputs into the engine as smaller chunks. Label this area:

Every interaction becomes a durable piece of knowledge.

3. Center: “The engine connects it”

Inside and immediately around the large central Chunk, depict a clear but elegant internal processing layer—not a generic AI brain.

Use three stacked, paper-card-like labels:

1. Extract & understand
   Titles, tags, content, source context

2. Index everything
   Semantic search across your
   Chunk

3. Connect the pieces
   Wiki-links · pins · collections · suggestions

Show small metadata marks, vector/semantic dots, and a subtle graph lattice emerging from the engine. Use a few concise technical labels in Spline Sans Mono styling:

• indexed
• linked
• grounded
• related
• retrievable

Avoid clutter. The system should look understandable at a glance.

4. Right side: “A searchable living knowledge graph”

Let the central engine expand into a beautiful, living knowledge graph on the right: cream, lake-blue, sage, plum, butter, and only occasional ember nodes connected with thin brown and ember lines.

The graph should contain recognizable object types:

• Notes
• Documents
• Research reports
• Conversations
• Captures
• Artifacts
• Collections
• Automations

Use varied node shapes—mostly rounded Chunk cubes and small paper cards—rather than abstract circles alone. Show a few explicit edge labels in tiny mono text:

• [[wiki link]]
• pinned to chat
• collected
• suggested connection

A highlighted search / retrieval card should sit over the graph:

Ask your whole Chunk

Find what you saved. See what connects. Use it when it matters.

Show a search field with an example query:

“What have I learned about spaced repetition?”

From the graph, a few relevant cards should surface toward the user: a note, a research report, flashcards, and a connected source. Indicate that answers are grounded in the user’s own knowledge.

5. Bottom outcome strip: “Knowledge compounds”

At the bottom, create a clean horizontal outcome sequence, with small Chunk icons between each:

Capture → Connect → Recall → Create → Learn

Supporting line:

Chunk turns the things you save, make, and discover into knowledge that keeps getting more useful.

Visual style and constraints

• Background: #FAF5EE warm paper, with extremely subtle paper grain.
• Cards: #FFFDF8, 1px warm ink border, 24px corner radius, soft warm shadows.
• Primary text:
  deep ink #2D2418.
• Secondary text: warm muted brown #6B5D4F.
• Primary accent: ember orange #E84D2B, used sparingly—only CTAs, active nodes, and the most meaningful connection paths. Keep orange below roughly 8% of the composition.
• Secondary accents: lake blue, sage green, butter yellow, and muted plum. Use them as soft tinted cards and graph-node categories.
• Typography: elegant warm editorial serif similar to Fraunces for the title and important labels; modern friendly sans similar to DM Sans for interface text; restrained mono styling for technical graph labels and wiki-link syntax.
• Shape language: rounded cubes/squircles, 16px buttons, 24px cards, 32px major panels. No oversized pill interfaces.
• Illustration style: premium flat paper-cutout + subtle 3D depth. No photography. No neon, glassmorphism, dark gradients, generic circuit boards, robot heads, or sci-fi server-room imagery.
• Tone: calm, intelligent, tactile, optimistic, rigorous. It should feel like a
  visual explanation of a product that helps someone build a connected second brain—not a generic AI diagram.

Key accuracy requirements

The diagram should communicate that Chunk’s knowledge graph is built from multiple durable object types—notes, documents, research, captures, artifacts, conversations, and collections—and that connections include wiki-links, pinned chat context, collection membership, and accepted suggestions. It should also show that Chat can retrieve and ground answers in those connected user objects.

This visualization is grounded in Chunk’s current capture, indexing, Connected Chat, durable edges graph, artifact, research, collection, automation, and memory architecture described in [[CLAUDE.md]] and [[Chunk — Strategy & Roadmap for the iOS 27 Era]].