How to use the matrix evaluation feature¶

The matrix feature enables systematic cross-product evaluation of items. It creates an interactive table where each cell represents the intersection of a row item and column item, which can be filled with AI-generated evaluations.

When to use a matrix¶

Matrices are useful when you need to evaluate combinations systematically:

Comparing business ideas against evaluation criteria
Analyzing features across different user segments
Rating products against multiple attributes
Evaluating options against decision factors

Creating a matrix¶

Select context nodes¶

Select one or more nodes that contain relevant context for your matrix
Type /matrix followed by a description of what rows and columns you want:

/matrix evaluate marketing channels (rows) against customer segments (columns)

The AI uses your prompt and the content from all selected nodes to extract appropriate row and column items.

Multiple context nodes¶

You can select as many nodes as needed to provide context. For example:

Select a node with a list of products
Also select a node with evaluation criteria
Also select a node with background research
Run /matrix compare products against criteria, considering the research context

All selected nodes become context for the matrix, and edges connect each context node to the resulting matrix.

Adjusting the matrix before creation¶

After running the command, a modal appears showing the extracted items:

Swap axes: Click the swap button to flip rows and columns
Remove items: Hover over any item and click the X to remove it
Review counts: The modal shows how many items are in each axis

Click "Create Matrix" when satisfied. The matrix node appears on the canvas, connected to all its source nodes.

Filling matrix cells¶

Fill a single cell¶

Click the + button in any empty cell. The AI evaluates that specific row-column intersection and streams the result into the cell.

Fill all empty cells¶

Click "Fill All" at the bottom of the matrix. A confirmation dialog appears showing how many cells will be filled. The AI processes each empty cell sequentially, streaming content as it generates.

Viewing cell details¶

Click any filled cell to open the detail modal, which shows:

The row item name
The column item name
The complete evaluation text (cells in the matrix show truncated previews)

Pinning cells to the canvas¶

When viewing a cell's details, click "Pin to Canvas" to extract that evaluation as a standalone node. This is useful when:

You want to reply to or branch from a specific evaluation
You need to reference a cell in further conversation
You want to include the evaluation in a different context

Pinned cells appear as dashed-border nodes connected back to the source matrix.

Practical example¶

Suppose you're brainstorming business ideas and want to evaluate them:

Ask the AI to generate a list of business ideas in one message
In another message, list your evaluation criteria (market size, competition, required capital, etc.)
Select both nodes
Run /matrix evaluate each business idea against these criteria
Review and adjust the extracted items
Create the matrix
Click "Fill All" to generate evaluations for every combination
Click cells to read full evaluations
Pin promising evaluations to the canvas for further exploration

Limits¶

Maximum 10 items per axis (100 cells total)
Cell content is truncated in the table view (click to see full text)
Fill All processes cells sequentially, which may take time for large matrices