future / semantic topology

What if meaning could guide layout?

Give the engine a page to analyze. It maps semantic affinity between content items, clusters what belongs together, and suggests an ordering — then tells you honestly whether the signal is strong enough to trust.

verdict promising clusters 3 anchor headline bridge pro-tier
I

The structure in your content

Choose a scenario. The engine embeds each item, measures pairwise affinity, clusters related modules, and renders a verdict.

Abstract semantic structure

Deterministic baseline

Embedding geometry looks strong enough to drive soft layout constraints.

promising
title Simple, transparent pricing 61 anchor
data Free plan 43 anchor
data Pro plan 59 anchor
data Enterprise 47 anchor
support Trusted by teams 45 anchor
cta Start your free trial 57 anchor
meta Common questions 34 anchor
separation 0.155 adjacency 0.646 centrality 0.569 query align 0.629 anchor headline bridge pro-tier
promising

Embedding geometry looks strong enough to drive soft layout constraints.

reading order headline → pro-tier → upgrade-cta → enterprise → social-proof → free-tier → faq-link
lead headline, pro-tier, upgrade-cta, enterprise Highest-intent cluster. Let this set tone and hierarchy.
body social-proof, free-tier Semantically close support group. Keep adjacency but give it more width.
support faq-link Lower-priority or utility information. Preserve access without crowding the lead.

headline reads as the semantic anchor: strongest blend of query relevance and centrality.

pro-tier behaves like the bridge: it has the highest cross-cluster affinity.

Clusters are meaningfully separated. Use them as spacing or zoning priors.

Workers AI not used. This report uses hashed token embeddings as a deterministic baseline.

The map above placed items where the math said they belong. The matrix below shows why — every cell is the cosine similarity between two items' embeddings.

Affinity matrix

Pairwise affinity between items.

Using the local heuristic baseline.

items
Simple, transparent pricing
Free plan
Pro plan
Enterprise
Trusted by teams
Start your free trial
Common questions
Simple, transparent pricing
1.00
0.49
0.60
0.44
0.33
0.67
0.51
Free plan
0.49
1.00
0.54
0.31
0.57
0.39
0.33
Pro plan
0.60
0.54
1.00
0.76
0.54
0.58
0.39
Enterprise
0.44
0.31
0.76
1.00
0.47
0.41
0.38
Trusted by teams
0.33
0.57
0.54
0.47
1.00
0.49
0.36
Start your free trial
0.67
0.39
0.58
0.41
0.49
1.00
0.37
Common questions
0.51
0.33
0.39
0.38
0.36
0.37
1.00
II

The break that follows thought

Same principle, narrower scope. Can sentence-level semantics improve where a paragraph breaks its lines — without reordering a single word?

sentences 4 clauses 8 avg similarity 0.546 recommended advanced
advanced

Pure width-aware typesetting already fits this paragraph well, so semantics add little.

Adjacent sentence shifts
related 0.51

The morning fog lifted slowly from the harbor, revealing masts and rigging in sharp silhouette.

We had planned to sail at dawn, but the wind betrayed us.

tight 0.77

We had planned to sail at dawn, but the wind betrayed us.

By noon the crew was restless, supplies were dwindling, and the barometer was falling.

weak 0.36

By noon the crew was restless, supplies were dwindling, and the barometer was falling.

Someone would have to decide whether to wait or abandon the crossing entirely.

Workers AI not used. Sentence and clause cohesion are derived from hashed token embeddings.

Advanced typeset

1039 badness

baseline
8 lines 361.34px max line 223.2px height
The morning fog lifted slowly from the
harbor, revealing masts and rigging in
sharp silhouette. We had planned to sail at
dawn, but the wind betrayed us. By noon
the crew was restless, supplies were
dwindling, and the barometer was falling.
Someone would have to decide whether
to wait or abandon the crossing entirely.

Optimizes rag, measure, and hyphenation only.

Syntax-aware

3611 badness

+2572
8 lines 366.3px max line 223.2px height
The morning fog lifted slowly from the
harbor, revealing masts and rigging in
sharp silhouette. We had planned to sail
at dawn, but the wind betrayed us. By noon
the crew was restless, supplies were
dwindling, and the barometer was falling.
Someone would have to decide whether
to wait or abandon the crossing entirely.

Rewards ending lines on punctuation and penalizes bridge-word lead-ins.

Syntax + semantic

6430 badness

+5391
8 lines 361.34px max line 223.2px height
The morning fog lifted slowly from the
harbor, revealing masts and rigging in
sharp silhouette. We had planned to sail at
dawn, but the wind betrayed us. By noon
the crew was restless, supplies were
dwindling, and the barometer was falling.
Someone would have to decide whether
to wait or abandon the crossing entirely.

Adds local clause cohesion so topic shifts prefer a fresh line.