Most SEO tools hand you a list and wish you luck. Seoluma measures what is actually happening to your site — rankings, technical health, link gaps, and whether AI engines cite you. RankMerit is the layer that turns those findings into the exact changes your site needs, and gets them made. Every draft goes through a gate built to reject it.
7.12B
links measured in-house
236.7M
domains in our own web graph
73
ranking pages scored to set our bar
We open with proof, not promises. This is a real crawl of your site, running now — the same measurement every engagement starts with.
An SEO operation has four jobs. Most tools automate one of them and call it autopilot. We cover all four, on one set of measurements, so a change in one is judged by its effect on the rest.
Meta titles and descriptions, heading structure, schema, internal linking, crawl errors, noindex accidents, broken canonicals, thin and orphaned pages. We find them, then produce the exact replacement — the tag to paste, the schema block, the redirect. Where you give us access, we make the change ourselves; where you don't, you get something a developer can apply in minutes.
Your keywords and your competitors, tracked on a regular cycle in the markets you actually sell in. Every movement is tied back to what changed on your site that week, so you can tell which action moved which position — and which did nothing.
Every article starts with the pages already winning that search: we read them, take them apart, and find what they all leave out. Then we write against your Search Console — the queries you get impressions for but no clicks, the positions one push from page one. Never to fill a publishing schedule: if there is no gap worth filling, we don't write a sixth version of the same page. Volume is not the product; each page has to earn its place.
How we write a page →ChatGPT, Perplexity, Google AI Overviews, Gemini and Copilot now answer instead of linking. RankMerit tracks whether they cite you or your competitor for your money questions — and restructures your entities, facts and sourcing until they cite you.
What GEO is →Most AI SEO tools take your keyword, hand it to a model, and publish what comes back. Same model, same day, every day. Here is what we do instead — and why it means we publish less.
| Tools that publish every day | RankMerit | |
|---|---|---|
| Where a page starts | A keyword and a prompt | The five pages already winning that exact search, in your country |
| Research before writing | None. The model writes from memory | Those five are downloaded and taken apart: structure, length, terms, data, sources, questions answered |
| What decides the structure | A fixed template, every time | What the winning pages actually do — we keep the shape readers and Google already accepted |
| What makes it worth publishing | It exists | It fills the gap all five leave open. If there is no gap, we do not write a sixth version |
| Checks before it goes live | None worth the name | Mechanical rules first, then a second, different model told to find a reason not to publish |
| How often it publishes | Daily, whatever the quality | Only what passes. Some weeks that means less |
| How it fails | 300 posts and no traffic | Fewer pages, each with a reason to exist |
What volume actually produces
We measured 70 tools in this category against our own web graph. One of the best-known daily-publishing tools has 1,357 sites linking to it — and 8 of them carry real editorial authority. Producing more did not earn more. It just produced more.
See the measurement →Method
Read the full writing process →
Seoluma is the measurement layer: its own web graph of 236.7M domains, real Google Keyword Planner volumes, rank tracking, and AI citation monitoring. RankMerit is its official partner — the layer that turns each of those signals into a change on your site.
Seoluma is our own product, not a vendor we license data from. That is the point: most SEO tools rent their link data and inherit whatever the supplier decides to measure. We built the measurement layer ourselves, so when it is wrong we can fix it — and when it finds something nobody else can see, it is ours to act on.
Anyone can export an audit and paste it into a chat window. What comes back is generic advice about a site the model has never measured, judged against an algorithm it last saw in training.
Connect up to three major AI providers to RankMerit and something different happens: they stop being observers. RankMerit hands each one the actual task — what to change, where, under what constraints, and against which measurement — written to the current state of the algorithm, then checks the result and reissues it if the number didn't move.
The models running your site aren't a stock assistant with a clever system prompt. They're trained for this work and driven by an algorithm written by practising SEOs — the rules that decide what to touch, in what order, and what to leave alone.
Connect your own models and RankMerit doesn't just show them your problems. It issues the task: what to change, on which URL, under which constraints, measured against what — versioned to the current state of the algorithm, not to what was true last year.
Every directive is executed, then re-measured against rankings and AI citations, then reissued differently if it didn't move. Pasting a report into a chat window gets you one answer. This is the part that has a next week.
01 / Authority data nobody else has
Most SEO tools rent their backlink data — or skip it entirely. Through Seoluma we run our own web graph: 236.7 million domains, 7.12 billion links. That's how RankMerit knows which sites actually pass authority in your niche, and which are recycled PBN noise.
02 / Answer engines, not just blue links
Most of the category is still optimising for ten blue links while a growing share of your buyers never sees a SERP. We measure citation share per engine per question, then fix the reasons you're skipped: unclear entities, unsourced claims, no extractable structure, no corroborating mentions elsewhere.
03 / No link buying. No spam. Ever.
RankMerit places no paid links, runs no directory blasts, touches no PBN. Good sites do not hand out body links to a guest post anyway — they cite sources. So the work goes into being worth citing. Our own Authority Index is that principle in practice: domain ratings for the entire web, computed from our 7.12-billion-link graph and published under CC BY 4.0 for anyone to check, quote or build on. Producing that kind of asset for your site is on the roadmap — and it will never be a link you paid for.
04 / A gate before anything ships
Whatever RankMerit produces is scored by a separate judge model whose job is to find a reason NOT to publish: unsourced statistics, cannibalised intent, missing entities the top-10 all cover. We calibrated it against 73 pages that already rank: their median score was 45, and not one reached 80. So the bar isn't a fixed number — your draft has to beat the best page currently ranking for that search. Below it, the draft is rewritten. Stubborn cases go to a human queue instead of your live site.
Setup takes about ten minutes. After that the only decision left is whether you want to approve changes or let it run.
See the full pipeline →Your domain, your Search Console, your target keywords. WordPress, Shopify, Webflow or a plain sitemap — it reads what you have in minutes.
Seoluma measures: technical state, rank positions, link gaps against a 7.12B-link graph, and whether AI answer engines cite you.
RankMerit does the work: fixes the technical faults, rewrites titles and structured data, adds internal links, and writes the articles and landing pages your Search Console shows are missing.
Each action logged against the rankings, impressions and AI citations that followed it. Not a vanity dashboard — a record of what changed and what it moved.
You don't buy a keyword list — you name the search terms you want to own. Each one is tracked daily in Google and in AI answers, built for, and revised until it moves. Three of those, run properly, beat a thousand sitting on a dashboard.
Connect up to three providers and you only pay for the platform: 1 site, 3 target keywords, daily SEO and GEO tracking, weekly on-page and technical auto-fix, and the quality gate. RankMerit drives your models with the directives — you just pay them for the tokens.
Every instruction sent to your models is logged, so you can see exactly what was asked and what came back.
what each target keyword gets
The moment you name a keyword it goes under daily measurement — its position in Google for your market, and whether AI answer engines cite you for it.
RankMerit builds what that keyword needs: the page if it's missing, the entities and schema if they're thin, the internal links pointing at it, and the technical fixes standing in its way.
Publishing isn't the end. Every piece stays measured in Seoluma — if it stalls at position 14, or the AI answers still cite someone else, it gets rewritten against what's actually ranking. Not abandoned for a fresh post.
Won a keyword? Retire it and name a new one in its place. Changed what you sell? Change all of them. On Agency, move keywords between client sites as priorities shift — the allowance is yours, not the site's. Everything already built for a retired keyword stays live on your site — and stays under measurement, so it keeps getting revised long after you've moved on to the next one.
or let us run everything
One site. Three keywords that matter.
3 target keywords$23.90/mo
For sites that need to compound.
8 target keywords$49/mo
Multi-market, multi-language.
20 target keywords$99/mo
Run it for your clients.
65 target keywords$249/mo
What actually happens after you pay
What you have to do
Approve things, if you want to. You choose whether changes are applied automatically or wait for your click. Everything we change stores the previous version, so anything can be undone.
If it is not for you
Ask for a refund within 14 days of your first payment and you get all of it back. No explanation needed, and the work already done stays on your site.
This is probably not for you if
Monthly. Cancel any time. Everything RankMerit produces is yours to keep — the drafts, the data, and the reasoning behind every decision.
No — and if you only need articles, we're the wrong purchase. RankMerit is an operation: technical repair, daily rank tracking, link-gap work and GEO. Content is one of four pillars, it is written against your Search Console demand rather than a schedule, and it's the pillar we gate hardest.
Generative Engine Optimization: getting cited inside the answers ChatGPT, Perplexity, Gemini, Copilot and Google's AI Overviews produce. Those systems don't rank ten links — they synthesise a single answer from sources they trust. Being on page one no longer guarantees you're in it. Different signals, different fixes, and it needs its own measurement.
We're not a customer of theirs — we're their official partner. Seoluma is the measurement layer, RankMerit is the execution layer, and the two were built to fit. Every signal Seoluma tracks arrives here as an instruction that can be executed, not a chart you're left to interpret.
Yes — up to three major providers at once. Connect your OpenAI, Anthropic and Gemini keys and the platform costs $9.90/month: measurement, daily SEO and GEO tracking, on-page and technical auto-fix, and the quality gate. RankMerit issues the directives; your models do the generation, on your account, and every instruction is logged for you.
For most sites, yes — and it's deliberate. These aren't rows on a dashboard: each one is a term we track daily in Google and in AI answers, build pages and entities for, and keep revising until it moves. Three of those, done properly, are worth more than a thousand keywords nobody acts on. And they're not locked in — change one, two or all three at the start of any month, and everything already built for the old ones stays live and stays maintained.
It stays under measurement in Seoluma. If it settles at position 14, or the answer engines still cite a competitor, RankMerit rewrites it against what is actually ranking — new entities, better sourcing, restructured for extraction. Most tools publish and move on; that's how sites end up with 300 posts and no traffic.
You'd get one generic answer about a site the model has never measured, judged against whatever it learned in training. RankMerit runs purpose-trained models under an algorithm written by working SEOs, gives them the specific task on the specific URL with the specific constraint, applies the change, measures what happened to the ranking and to the AI citations, then reissues it differently if nothing moved. A screenshot has no next week.
Never. Paid link schemes violate Google's link-spam policy and are increasingly cheap to detect. We build assets worth citing and put them in front of the people who cite that kind of thing. Slower — and it doesn't evaporate at the next update.
No. Nothing goes live on your site without you seeing it first. Every change is prepared, shown to you with what it replaces and why, and applied only once you approve it — by us where you have given access, or by your own developer from what we hand over. Everything we change is logged with the previous value, so any of it can be put back.
Technical fixes land in the first week and often move impressions within days. Content, authority and GEO work compound over 60–90 days — anyone promising rankings in two weeks is selling you something else.
One minute, no signup, no card. We'll crawl your pages, pull the real search volumes and tell you exactly what's standing between you and the traffic.
Scan my site — free