Semantic Core · Satellite1
← Back
● semantic-core · april 2026 · v1.0
RU
● pbn-workers.com / semantic core

Not a keyword list — a ready-made site architecture.

We build your semantic core turnkey: from a short topic brief to cluster structure and sections. You get a content map — what to write, in which section, targeting which query.

up to 500,000+
keywords in a single core
from 24 hours
vs. 2–4 weeks by hand
from ₽1
per keyword at volume
Sound familiar

What people come with.

"A contractor handed me 5,000 phrases in Excel — now what do I do with them, where do they even go?"
"We publish articles on a hunch. Budget goes into content, but rankings don't move."
"I don't understand which pages to build and how they should link to each other."
What you get

You get an architecture, not a word table.

Every order is not a parser export but a ready-to-use package: site structure, clusters, and a map of how queries are distributed across pages. Below is what a single cluster looks like in a real core.

01
Clusters tagged by intent
Marker keyword, search intent (commercial / informational / navigational), cluster size and total search volume.
02
Breakdown by search volume
Keywords in each cluster split into high-, mid- and low-volume — so you see what to fight for first.
03
Silo structure of sections
A hierarchy of sections with descriptions and nesting — a ready site architecture, not a flat list.
04
Query-to-page map
Which cluster maps to which page and section — a ready distribution the site structure is built from.
05
Format of your choice
Spreadsheet or JSON — whichever slots into your current workflow more easily.
● fragment of a finished core · one cluster out of several hundred
format: spreadsheet / JSON
cluster_02distance learning
informational intent
218
keywords in cluster
50,449
total search volume
query exact broad
what is distance learning
6,498
33,528
part-time study format explained
5,756
11,705
distance learning meaning
1,411
6,597
how distance learning at university works
717
1,471
difference between full-time and distance learning
144
1,280
+ 213 more queries in this cluster, split into high-, mid- and low-volume
This is what a single cluster looks like: marker, search intent, volume and frequency — plus every query inside with its exact and broad volume. A finished core holds hundreds of such clusters, wired into a section structure.
● Silo structure · how clusters add up to a site architecture
section Study formats
subsection Distance learning
cluster what is distance learning 218 keywords · 50,449
page /distance-learning
subsection Part-time format
cluster part-time learning 164 keywords · 38,900
page /part-time
Each cluster lands on its own page, subsections become site sections, everything wired into a tree. That's the ready architecture: a developer or content manager just builds pages from the finished map, instead of inventing which sections to create and how to nest them.
Which use cases

For a live site and for a project from scratch.

People come with different goals. We'll build the semantics for a specific type of site, or hand over the structure of an entire project before development even begins.

existing site
Blog / content project
We map topics and sections across the whole demand in your niche. You'll know what to write for months ahead and how to link the material together.
existing site
Catalog / online store
We break demand down across categories and product pages and find the blank spots — sections you don't have yet where demand already exists.
existing site
Services site
Clusters for each service and related queries, with commercial and informational pages clearly marked.
● from scratch
A whole new project
No site yet — we hand over the semantics and structure of the entire project before development starts. The site is built right from day one, with no structural rework later.
Why it works
● the key difference
We keep one intent in one cluster — even across tens of thousands of near-identical queries.
At high volumes of similar semantics, automated tools break down: queries with the same intent scatter across different clusters, and you then have to merge them by hand — slow and expensive. A specialist can only pull this off manually on a page or two, not across a whole project. We run several clustering methods and pick the one that fits your SERP without splitting intent — so scale into thousands and tens of thousands of queries doesn't fall apart.
●
Google and Yandex at once
We gather demand and group the SERP across both search engines instead of favoring one. The core works in both — critical for bilingual and multi-engine markets.
●
Built in-house
Clustering runs on our own algorithms, not a rented service. We control the grouping logic and pick the method per niche, instead of bending to someone else's limits.
●
Morphology-aware
"buy", "buying", "bought" — one keyword, not three duplicates in a cluster. Full morphological analysis of word forms.
●
Niche map
Before collecting, we build a map of subtopics and angles — systematic coverage of the topic, not random phrases around the obvious.
●
Junk filtering
We isolate geo-modifiers, flag aggregator queries, and cut the short and junk tail — before it ever reaches your core.
●
Reproducibility
The same set of keywords gives the same result on a re-run. The semantics can be refreshed as the site grows.
Pricing

Price per keyword drops with volume.

Base rate is ₽5 per keyword. The larger the core, the lower the price per query. Get a ballpark below — the final price is calculated after collection and cleanup, and you pay only for approved keywords.

Price per keyword
₽5.0
base rate
Project estimate
₽5,000
1,000 keywords
Turnaround
from 24h
after SERP collection
Figures are a ballpark. Exact volume and price are set after collection and cleanup; you pay only for approved, clustered queries.
Retail volumes
VolumePrice per keywordDiscountProject estimate
up to 1,000₽5.0—from ₽5,000
1,000 – 5,000₽4.5−10%~₽22,500 for 5K
5,000 – 10,000₽4.0−20%~₽40,000 for 10K
10,000 – 30,000₽3.5−30%~₽105,000 for 30K
30,000 – 50,000₽3.0−40%~₽150,000 for 50K
Wholesale volumes · aggregators, marketplaces, multi-region
VolumePrice per keywordDiscountProject estimate
50,000 – 150,000₽2.5−50%~₽375,000 for 150K
150,000 – 300,000₽2.0−60%~₽600,000 for 300K
300,000 – 500,000₽1.5−70%~₽750,000 for 500K
500,000+₽1.0−80%custom quote
We'll size up your volume and send a sample core.
Free, under NDA, no strings attached.
@antonivanov89 · consulting
Comparison

Price, turnaround, quality — against the market.

Reference point — a 10,000-keyword core. Freelance is cheaper, but it's an automated export that gets reworked. Agencies are comparable on price, but they collect over weeks and by hand. The gap shows clearest at high volumes of similar semantics — where tools split intent and a specialist can't merge it all by hand in time.

Parameter Freelance marketplace Auto-service SEO agency Satellite¹
Price (10K) ~₽1K ~₽5K ₽30–85K ₽40K
Turnaround 2–5 days hours 2–4 weeks from 24 hours
Clustering basis manual / varies by SERP, one algorithm by SERP by SERP, several methods
High volume of similar queries only for 1 page splits intent manual rework holds structure
Cluster intent tagging — partial yes ✓ tagged
Silo structure — partial manual ✓ built
Query-to-page map — — manual ✓
Morphology handling hit or miss partial depends full
Reproducibility — partial — ✓
Under the hood
Technical details for the curious expand ▾
Several clustering methods
We pick the algorithm to fit the job: from a fast balance of speed and quality, to strict grouping with minimal junk overlap, to a hybrid mode for broad commercial niches full of similar queries.
SERP-based grouping
Two phrases land in one cluster if the same pages rank for both. Grouping reflects how the search engine itself reads the queries, not word overlap.
Morphology handling
Analysis and dedup by the base form of each word. Word forms collapse into a single keyword, so clusters aren't cluttered with duplicates.
Reproducibility
A re-run on the same keyword set gives the same result and computes noticeably faster. The semantics are easy to refresh as the site grows.
What we need from you

One brief — we handle the rest.

All we need is a 2–5 sentence topic brief: what the site sells or writes about, the region, the type (online store, blog, services). Everything else is on us.
Optional, if you have them: marker queries stop words target region
Questions

What people usually ask.

What if the niche is narrow and there aren't many keywords? ＋
For narrow topics the core is smaller — and that's fine. You pay for the keywords actually collected and approved, not for a forced quota. We won't inflate the core with irrelevant queries: extra phrases bring the wrong people to the site.
For Google or Yandex? ＋
Both. We gather demand and analyze the SERP of both engines, so the structure and priorities account for Google and Yandex alike. If you're only targeting one — tell us and we'll focus on it.
Can I review and adjust the result? ＋
Yes. You approve the makeup of the core before final delivery and pay only for the clusters you accept. Need to fix the tagging or drop some queries — we make the edits before export.
How accurate is it compared with manual work? ＋
Grouping runs on the real SERP, not on subjective judgment. The result is reproducible: the same keyword set on a re-run yields the same structure — unlike manual work, where the outcome depends on the individual specialist.
What about confidentiality? ＋
We work under NDA. We don't show or reuse your project's topic, keywords, or structure. The finished core stays yours.
What's next
● after the core
We can fill the ready structure with content.
The core is a map. From there we write articles against it with our own methods: not "text on a topic," but a page built for the real SERP behind a query. Along with the content we hand over page skeletons — topic, section and keywords for each. Every page runs through the same pipeline with a check at every step — moving to the next stage only after the previous one clears the quality thresholds.
01
SERP snapshot
We break down the top for the query: what the ranking pages cover and how.
02
Page semantics
The terms and entities the search engine expects in text on the topic.
03
Structure and copy
Headings, emphasis, facts — to one editorial style across the project.
04
Pre-delivery check
Uniqueness, no keyword stuffing, no fabricated facts.
05
Refresh
We update the page after publishing — the search engine sees freshness.
● Every stage is checked against numbers. Until a step passes its check, we don't move to the next.
● an article for every cluster in the core
● for Google and Yandex at once
● no filler, no keyword stuffing
We'll send a sample finished core for your niche — free.
Describe your topic in a couple of sentences — on an NDA consultation we'll show the structure, the numbers, and size up your volume. No strings attached.
consulting@antonivanov89
channel@satellite1ru