Architect / Mind
/mindis the surface where Architect's memory layer becomes navigable. Locked sessions, anchored asks, promoted canvas tiles, imported citations — every unit is a node, every concept-overlap a thin line of light, every reading position a different gravity well.
/mind · Exploring · architect.eleven11.pro
Reading the screen
Nothing in /mindis decorative. Each region of the screen is a window into the same memory layer — what shifts is which slice you're looking at.
01 · Left rail
Every memory unit your workspace has produced — locked sessions, anchored asks, promoted canvas tiles, citations imported from Granth, references and articles. Sorted, filterable, searchable. The list is the long version of the graph.
02 · Center node
The brightest node is the unit you're focused on. Selecting any unit on the left pulls it to the center; the graph rebuilds around it. The graph is not a static map of your workspace — it's a reading position.
03 · Radiating edges
Each line is an edge between two memory units. Brighter, thicker edges mean stronger conceptual overlap. The shape isn't decorative — the substrate computes overlap from your corpus alone, and the visualization mirrors what recall would surface if you asked the workspace a question right now.
04 · Search + filter
Search resolves to a focal node and rebuilds the graph around it. Filters narrow by kind, matter, date, or edge type — so a research-heavy session and an operations runbook draw different graphs from the same substrate.
What the edges mean
The graph is not a guess. Every edge is computed from your own units, and every edge carries provenance — the block_id that produced it, the action that approved it, the timestamp when it landed.
concept-overlap
Computed from your corpus. Two units share enough of the same concepts that recall would surface them together. Most edges in /mind are this kind.
citation
An ask, a session, or a Granth document explicitly cited a unit. Provenance is preserved — every citation edge carries the source block_id.
lineage
A unit was promoted from another — a canvas tile became a session, a session became a matter section. The chain is preserved as an edge.
manual
You linked them. Suggestions become edges only after you accept them; the graph never accumulates connections you didn't see.
Three things you do here
The verbs are deliberately small. /mindis a reading surface, not a workspace. What you find here goes back to a session or a matter — that's where work happens.
01
Type a phrase. The graph resolves to the closest unit and recenters. Pivot to any neighbor in two clicks — drill, peek the body, jump.
02
Click any node. The right panel opens its body, citations, and outbound edges. Open a neighbor and the graph rebuilds around the new center.
03
Pull any unit back into a live session as a citable block — or anchor it to a matter as a memory reference. The mind is read-only by default, but it never traps what you find there.
Mind and memory
Memory is the layer. Mind is the surface. The units you see in /mind are the same rows that ground every model call in Architect — every ask, every demand, every locked session.
What changes between the two is the angle of attack. Memory is what the substrate uses without you noticing. Mind is what the substrate looks like when you decide to look.
Read the memory explainer for the five promises the substrate makes — /architect/memory.
Open /mind
