Chapter 4: Mindware
Mindware is a platform for packaging concept structure models as content, publishing and selling them on the store. You can purchase models built by other users, explore them, and deepen your knowledge through AI conversations.
4.1 Screen Layout (3 Tabs)
Click “Mindware” in the sidebar to see three tabs at the top.
Store Tab:
- List of published Mindware content
- Displayed in card grid format
- Filterable by search bar and category filters
My Library Tab:
- Purchased or freely accessible content
- Your own Mindware also appears here
My Models Tab:
- Register and manage your concept structure models (CCM groups) as Mindware
- Available for Personal plan and above
4.2 Content Card Layout
Each Mindware content item is displayed in card format.
| Element | Description |
|---|---|
| Cover Image | Visual identity of the content. Accent color gradient if not set |
| Title | Content name |
| Author | Creator name |
| Description (1-2 lines) | Content summary |
| Price Badge | “Free” / “$XX” / “Purchased” |
4.3 Content Detail View
Click a card to open the detail page.
Detail Page Layout:
- Hero image (large cover image)
- Title and author name
- Full description
- List of connected models (which text columns they were built from)
- Source information (books, papers, etc.)
- Categories and tags
Action Buttons:
- “Purchase” button (paid, not yet purchased): Redirects to Stripe checkout
- “Open” button (purchased/free): Opens the Explorer view
- “Back” button: Returns to the store list
4.4 Purchasing Content
- Click a content card in the store
- Review the content on the detail page
- Click the “Purchase” button
- Stripe checkout opens in a new tab
- Enter credit card information and complete payment
- After checkout, you return to ThinkNavi
- Content is added to “My Library”
Subscription-based Content:
Some content uses monthly billing. It auto-renews until canceled.
4.5 Exploring Models in the Explorer
Opening content takes you to the Explorer view.
Network View
Displays the concept structure using an SVG-based 2D graph.
Display Elements:
- Nodes (circles): Concept representative points. Color-coded by cluster. Size proportional to data volume
- Edges (lines): Connections between nodes (MST)
Operations:
- Click a node → Node detail panel appears
- List of assigned text chunks
- Summary of each chunk
- Values for each dimension
- Cluster ID
Exploration View
Statistical list view of clusters and nodes.
Display:
- Cluster list
- Expand a cluster → Top texts (summaries of representative data)
- Node list and their data assignments
Source Text Context View
View the original text of data assigned to a node with surrounding context.
Operations:
- Click a node to open the detail panel
- Click the “View Context” button for assigned data
- Previous 2 chunks — Current chunk (yellow highlight) — Next 2 chunks are displayed
4.6 AI Chat (Agent)
AI chat is displayed in the right panel of the Explorer. It enables conversations leveraging the model’s knowledge.
Operations:
- Enter a question in the text input field
- Send to have the AI search the model and respond
Context Integration:
- Select a node or cluster in the Network View and click the “Ask AI” button
- Selected context information is automatically inserted into the chat
AI Response Principles:
- AI directly quotes original text from nodes in the model
- AI does not use its own knowledge or interpretations
- If no relevant information is found, AI honestly states so and suggests related topics within the model’s scope
- Source node IDs are included in responses
Tools Used by AI Internally:
search_nodes(query, model?): Search nodes by text queryget_node_detail(node_id, model): Get full node detailsget_cluster_details(cluster_id, model): Get cluster dimension profile and member nodesget_table_data(filter?, columns?, limit?, offset?): Search the original data table
Note: AI may convert questions to English before searching (if the model data is in English). Searching works correctly even when asking in Japanese.
4.7 Dialektike (Discourse Dojo)
A gamification feature for deepening understanding of model concepts through Socratic intellectual dialogue.
Starting Discourse Dojo
- Click the scales button in the Explorer view
- AI chat switches to “Discourse Dojo Mode”
- Click the “New Question” button
Flow of Discourse
- AI selects a random text from the model and presents an intellectual question based on it
- User responds freely
- AI acts as an equal dialogue partner:
- Presents its interpretation with evidence (text citations from the model)
- Counters weak arguments with specific counterexamples
- Poses exploratory questions about conceptual boundaries
- Through repeated dialogue, deepen understanding of concepts
Session Management
- Click the “History” button to view past discourse sessions
- The first question and answer of each session are shown as preview
- Select a session to reload it
- Unnecessary sessions can be deleted
4.8 Creating and Publishing Mindware
Users on Personal plan or above can publish and sell their models as Mindware.
Registration Steps
- Click the “New Registration” button on the “My Models” tab
- Fill in the registration form:
| Field | Required | Description |
|---|---|---|
| Title | Yes | Content name |
| Description | Yes | Content summary (displayed in the store) |
| Author Name | Yes | Displayed author name |
| Category | Yes | Content category |
| Tags | Comma-separated keywords | |
| Sources | Information about books/data used | |
| CCM Group | Yes | Link to a model (built with ConceptMap-Text) |
| Accent Color | Card background color | |
| Cover Image | Drag & drop or file select (max 10MB) | |
| Pricing | Yes | Free / One-time purchase / Subscription (monthly) |
- Click the “Save” button
- For paid content, the product and price are automatically registered with Stripe
Cover Image Specifications:
- Maximum file size: 10MB
- Supported formats: JPEG, PNG, WebP, GIF
- Automatically resized and converted to WebP format on upload
Editing and Deleting
- Hover over a card in the My Models tab to see edit and delete buttons
- Confirm with the “Delete?” dialog
Note: Deleting paid content does not revoke access for users who have already purchased it.
4.9 Troubleshooting
| Issue | Cause and Solution |
|---|---|
| “Open” button is grayed out | CCM group is not linked to the title. Select a CCM group in the edit screen |
| Purchase completed but not showing as “Purchased” | Reload the page. Stripe Webhook processing may take a few seconds |
| First AI chat response is slow | Engine session rehydration (reconstruction) is occurring. The first chat takes 10-30 seconds, but subsequent responses are fast |
| Image upload error | Check that the file size does not exceed 10MB and is a supported format (JPEG/PNG/WebP/GIF) |
| “Please select an image file” alert | A non-image file was selected as the cover image. Please select an image file |