Home/Features/Keyword Lattice
Keyword LatticeMap

Keyword research grounded in your own pages

Most keyword tools hand you a list with no connection to your site. This one starts from your own pages and the SERPs they compete in, so every keyword arrives already attached to the section of the site that has to rank for it.

quietly.sh/keywordsscreen recording
The keyword constellation under a slow zoom: 120 keywords resolved into 12 clusters, each hub labeled for the part of the site it covers.

What it does

Keywords your site and its SERPs actually use

Three sources, all of them yours: the repeated phrases in your page copy, the phrases in your page titles, and the wording of the top ten organic results for your seed. Phrases carried by most of your pages are dropped as site chrome, so navigation labels do not come back as opportunities.

Grouped by the part of the site that owns them

Clusters follow your URL structure. Everything under /blog becomes one hub, /case-studies another, the homepage its own. Each carries a coverage score, the average semantic match between the pages in it and your seed, so a thin section is visible as a low number.

A score that tells you what it did not know

Priority combines search volume, difficulty, your current rank and how well your existing copy matches. Any input it does not have is left out of the average instead of counted as a zero, and every keyword carries the confidence that produced it: low, medium or high. A score built on one measured input says so.

Coverage against your competitors, labeled as coverage

Point it at rival domains and it reads the phrases in their indexed pages, then shows which ones write about each phrase and which do not. That is topical coverage, and the column header says so. Rank comes from the tracker.

Every keyword has somewhere to go

From a keyword: start tracking its rank, generate a content brief, open a takeover analysis of who holds the SERP, run a content-gap pass, write the article, or send the whole cluster to Fabric as rows on a schedule.

quietly.sh/competitorsscreenshot
Twenty-five rivals derived from 122 tracked keyword SERPs, with the queries each of them leads on: reddit.com is ahead on 20 of them at an average position of 2.7.
Twenty-five rivals derived from 122 tracked keyword SERPs, with the queries each of them leads on: reddit.com is ahead on 20 of them at an average position of 2.7.

How it works

  1. 1

    Give it a seed and a site

    One phrase and the site it belongs to. It crawls the pages, reads their copy and titles, runs the seed and two variants through live search, and returns around 120 keywords per run.

  2. 2

    Read the constellation

    The constellation shows the hubs and what each one covers. Intent is classified from the wording, so pricing and buy land in transactional and how-to lands in informational. A phrase whose wording says nothing is marked unknown.

  3. 3

    Turn one into work

    Track it, brief it, write it, or push a cluster into Fabric where each row runs on its own schedule.

On search volume

Volume and difficulty come from DataForSEO when you connect an account, and they are blank when you have not. Nothing here estimates them locally, because a made-up volume figure is worse than an empty column: it looks like evidence. Everything else on the page, the clusters, the coverage comparison, the rank checks and the priority ordering, works without it.

Find the keywords you are missing

Seven-day trial. Point it at a site you own and read the first constellation in a sitting.