Backed by Y CombinatorFour documents and four sheets that score a cluster's link graph before any brief goes out, not after the drafts come back.
Free download · No account needed
What you get
Pack of 8 files: 4 documents, 4 spreadsheets
Pillar Page and Topic Cluster TemplateLink Plan
Generated from the adjacency map on Cluster Structure, after Coverage Gaps sorts the keyword map.
| Source page | Source section | Anchor text | Target page |
|---|---|---|---|
A row with no Source section is the one that never gets written.
Every pillar page template on page one ships the same diagram: spokes link up to a pillar, the pillar links down to the spokes. A star. What the diagram omits is what a star actually gives each spoke, which is exactly one inbound link from inside its own cluster, the pillar's. Modelled across a 222-page site, that topology leaves the worst-off spoke holding 0.245% of the site's internal authority, the same share whichever fourteen posts you pick.
This pack scores the graph before any brief goes out. Coverage Gaps sorts the keyword map first, so a query that is really a section of another page never becomes its own thin post; on the worked cluster seven of twenty-two queries turned out to be sections, not spokes. Cluster Structure then gives each spoke exactly three sibling links, chosen for a reason stated in one sentence. The Link Plan turns that map into one row per link, naming the exact section of the source page it lives in.
The result: modelled at the same fifteen pages, the planned graph raises the cluster's total internal authority 43.8% and the worst-off spoke's share 152.7%, with no new content and no new external links. Feed it a completed keyword clustering map and whatever already exists, or send River both and let it sort role, adjacency and anchor text before a single brief goes out, each one already carrying the links it owes.
Coverage GapsFictional company: Marlowe Ops. 22 queries sorted before anything is commissioned.
| Query | Role | Existing URL | Action |
|---|---|---|---|
| shift scheduling | pillar | none | write the pillar |
| rotating shift patterns | spoke | 2 URLs | merge the pair |
| employee scheduling app | section only | none | no page |
| time tracking for shift workers | wrong cluster | /blog/shift-time-tracking | reassign it |
Seven of 22 queries are sections of a page, not pages of their own. That is what keeps the cluster at 15 rather than 22.
Cluster StructureOne row per page. Links out and the in-degree target are set before any brief is written.
| Page | Role | Links out to | In-degree target |
|---|---|---|---|
| /shift-scheduling | pillar | all 14 spokes | 14 |
| /blog/split-shift-rules | spoke | overtime, differential pay, predictive laws | 3 |
| /blog/on-call-scheduling | spoke | split shift, coverage, swapping | 3 |
| /blog/shift-handover-checklist | spoke | build, software, overtime | 3 |
Every spoke links to the pillar and to exactly three siblings, each with a reason stated in one sentence.
Link Graph ComparisonSame 15 pages, five link plans, scored by modelled PageRank share.
| Plan | Cluster total | Worst spoke | Sibling links in |
|---|---|---|---|
| Star, pillar in blog index | 10.746% | 0.245% | 0 |
| Star plus two laterals | 13.704% | 0.374% | 2 |
| Planned, floor of three, plus nav link | 15.448% | 0.618% | 3 |
Cluster total +43.8%, worst-off spoke +152.7%, on the same fifteen pages with no new content.
Sorts every query into a page, a section of a page, or the wrong cluster, before anything is commissioned.
Four rules a link has to satisfy before it counts, including the anchor-plus-href test Google's crawlers actually apply.
The adjacency map: every page's role, its links out, and the reason each connection exists in one sentence.
Scores five link plans on a modelled 222-page site, so a floor of three links is arithmetic, not a guess.
The scored result: a planned graph raises cluster authority 43.8% over the star every template produces.
One row per intended link, each naming the exact section of the source page and the anchor text it uses.
Fourteen handoff sections in adjacency order, each ending in the exact sentence and anchor the Link Plan assigned it.
One per spoke, with the links it owes, the links it receives, and the outline built from both.
Send the keyword cluster map and whatever already exists. Coverage Gaps sorts each query into a page, a section, or the wrong cluster.
Each spoke gets exactly three sibling links out, chosen for a reason stated in one sentence, with an in-degree check on every page.
The adjacency map becomes one row per link, naming the exact section of the source page and the anchor text it carries.
Pillar and spoke briefs generate from the finished plan, so every outline already has a place for the links it owes.
Yes, no account or card is required for the Word and CSV download. Edit with AI installs this exact pack in River and reads your keyword cluster map before sorting a single query. A spoke that stops ranking is a different job, covered by the content refresh pack.
A related-posts widget is written by a script after the page loads, and Google's own guidance says most links in other formats will not be parsed by its crawlers. A conventional cluster template diagrams a star, where every spoke's only inbound link is the pillar's. This pack plans anchor-in-prose links with an in-degree floor of three, scored first.
Google's own documentation sets the floor at one: every page you care about should have a link from at least one other page. Three is what the arithmetic pays for. Going from the pillar's link alone to three sibling links moves the worst-off spoke from 0.245% to 0.618% of the site's modelled internal authority, a factor of 2.5, on the same fifteen pages.
It generates the briefs, not the prose. Once the plan is finished and checked, Write the Pillar and Spoke Briefs turns it into an outline with a named section for every link a page owes. A writer, or an SEO writing review, then has something concrete to check the draft against.
This pack sorts and links a map that already exists; building one from scratch is a keyword research job first. A keyword clustering map, even a rough export from any keyword tool, is the one input the space cannot start without.
Fewer than the keyword map suggests, usually. On the worked cluster twenty-two queries sorted down to fifteen pages, because seven turned out to be sections of a page rather than pages competing for their own ranking. Coverage Gaps runs first specifically to catch that before anything is commissioned.
Written by
Co-founder & CEO, River
Chandler has spent the last six years working in AI, building tools for millions of users and Fortune 100 companies. In 2025, he co-founded River to build a new kind of operating system for AI agents.
Send the keyword cluster map and whatever already exists. River sorts, plans the adjacency, and generates every brief with its links already placed.
Edit with AI