← Back
Manual — AI & SEO Guide

rankmio Manual

Updated

Everything you need to get the most out of rankmio — step by step.

1Getting Started

rankmio combines technical SEO analysis, AI-powered content optimization, and Generative Engine Optimization (GEO) in a single workflow: Analyze → Optimize → Verify → Report. Instead of ten tools, you only need one.

You create an account in a few steps

1
Register at /register with your name, email, and password.
2
Confirm your email address via the verification link.
3
You receive 10 free credits to test all features.

Your first project needs only the domain

1
After login, you will be prompted to create your first project (= a domain).
2
Enter your domain (e.g. example.com). rankmio creates the start URL and checks whether its crawler can reach the website; if a protection service blocks it, a notice explains what to do (see chapter 9).
3
Optional: Connect Google Search Console under Connections → GSC to get keyword data, clicks, and impressions. No GSC yet? See the setup guide.
4
Start your first website analysis — results in about 30 seconds.
Tip: Connect GSC as early as possible. Many features (Opportunities, Trends, Cannibalization, Rank Alerts) require GSC data as a foundation.
The rankmio Wizard guides you step by step through nine typical use cases.

The free check works without registration

With the free SEO check you can analyse any URL without registration: a weighted SEO score (0–100) and the measured AI citation readiness, that is how well the page itself is prepared to be cited by an AI. Without login, one check per IP address and day is free. With login, you can run as many checks as you like, each one costs credits. Results remain visible for one year.

2Dashboard

The dashboard is your project cockpit. At the top it shows the state of the project in four figures, below that the tasks and the blocks in which there is actually something to do. Everything always refers to the project selected at the top.

Four figures summarise the project at the top

The large ring shows the visibility index from 0 to 100 with the change against the previous week. Beside it, the Analysis status button opens the panel that shows which analyses have already run for this project; it is described further down under Analysis status and rankmio Score.

Below the ring there are three tiles. The SEO checklist shows the share of passed items from the technical check. The GEO score is half citation readiness and half actual citations: citation readiness counts which prerequisites your pages fulfil so that an AI can cite them, actual citations measure how often AI systems really cite you. Technical health combines load time, Core Web Vitals and the crawler findings. Each tile shows the change against the previous week, and a click opens the breakdown.

If you select a single page at the top, the tiles show its values wherever a value per page exists, and they say so where a figure only exists for the whole project.

The blocks below show where there is work to do

Top 5 fixes

The five most urgent tasks of the project. The complete list is under Actions, the audit with period comparison opens from here.

Project tasks

Tasks that concern the whole project, such as the sitemap or the robots.txt. They remain visible whichever page you select.

Progress hub

Your metrics over time: rankings, clicks, impressions and AI visibility. A sentence below the metrics tells you how your existing search queries on pages 1 and 2 moved, separately from new and dropped ones. The Rankings tab adds the figures.

Visibility index

The four sources of the index individually, with the ceiling as long as a source is missing.

Keyword quick wins

Search terms in positions 4 to 10 with enough data where work pays off soonest. Details opens the opportunities in the Keyword Hub, Optimise jumps straight to the query there.

These pages are losing clicks

The three largest click losses per page in the last 28 days compared with the 28 before, each with the query that lost most. A loss counts from two clicks; pages with fewer than ten clicks before are left out because chance is possible there. Details leads to Traffic recovery.

AI search status

How often your domain is cited in the AI systems your citation trackers query.

Competitor comparison

Your own domain next to the pinned competitors.

The tiles above the trend follow the period you choose

Above the trend in the progress hub there are three tiles: Average position, Clicks and Impressions. They cover the period you choose for the trend, that is 7 days, 28 days, 90 days or 1 year, and each tile names it. When the dashboard opens, 28 days apply.

Only days with final data count, and the position is weighted by impressions. At 28 days the tiles therefore show exactly the figures of the Keyword Hub. Bing counts towards clicks and impressions as soon as Bing data exists for the period, and the tile shows the split, for example “57 Google · 4 Bing”.

The change compares with the equally long period before. If the data is not sufficient for that, the tile says in one sentence from when the comparison will be possible.

The AI visibility tile counts Copilot citations from Bing over the last 28 days for which Bing data exists.

All Search Console and Bing figures cover the same 28 days

The dashboard, Keyword Hub, tasks, chat assistant, the MCP interface for AI agents and the JSON export all use the same period: 28 days up to the last day for which Google delivers final data. This matches the default “Last 28 days” view in Search Console. The period covers four full weeks, so every weekday occurs equally often.

Comparisons between now and before set these 28 days against the 28 days before. Bing figures end on the last day for which Bing data exists. Only where you choose a period yourself, such as in the trend of the progress hub, does your choice apply.

Until 8 November 2026, some views calculated with 30 days counted back from today and others with 28 days. The same page could therefore show two different figures in two views.

rankmio collects every finding in one task list

Every analysis reports what it finds, and rankmio turns it into tasks. The same defect on the same page appears only once, even if several analyses report it. The dashboard, the Actions list and the JSON export show the same list.

Each task carries a level that says how urgent it is.

LevelMeaningExamples
CriticalThe page cannot be found or used.Error code, noindex, robots.txt block, no HTTPS, no title, unusable on mobile
HighA concrete lever or a proven risk, but nothing is broken.A defect on most pages, a page Google does not index
MediumThe page works but leaves a measurable lever unused.Missing description, missing schema, images without alt text, long load time
NoteMeasurable, but the effect is small or unproven.Several H1 headings, which Google expressly allows

In addition, each task names what it affects: crawling, indexing, ranking, click rate, AI visibility, usability or technology. "Critical, indexing" tells you more than a red icon.

In the Actions list you can filter by level, by topic and by finding. If at least five tasks of the same kind are in the list, a combined task takes their place; the individual pages remain visible in the respective report. Bundling is switched on from the start. You can switch it off in the Actions list by unticking “Bundle similar tasks” or in the dashboard with “show individually”. The setting is saved and also applies to your AI agent.

Tasks that apply to the whole project are listed separately from those of individual pages. If you mark pages as legal, function, tool or directory pages under Page roles, tasks that do not fit their purpose disappear; errors remain visible.

When you tick off a task that can be checked on its page, such as a missing title or a missing description, rankmio offers to fetch the page again. The check costs no credits. Your tick holds until the next measurement: if the next crawl, GEO audit or Search Console sync still finds the defect, the task reopens and carries the note “Still present despite being ticked off”. Tasks without a measurement of their own, such as those from the AI action plan, stay ticked off. Each task also names the date of the measurement it is based on, and when a new measurement no longer finds a defect, the task disappears by itself.

For cannibalisation, the task recommends one page per topic

If several of your pages compete for similar search queries, one task is created per topic. The topic consists of the distinguishing words of the query: all queries about “Sistrix alternative” belong together, while a comparison such as “sistrix vs seobility” is a topic of its own. At most five topics appear, sorted by leverage.

The task lists all pages involved in a table with queries, impressions, clicks and position, and recommends one of them. Clicks decide first: the page with the most clicks wins, provided it has at least 3. Otherwise the best position decides, and a page that carries the term in its address or title may be up to five positions behind and is then preferred.

For the other pages, the task says how to defuse the competition: take the term out of the title and headings, shorten the section and link to the recommended page using the search term. If at least 80 per cent of the impressions of a page come from the topic, merging it into the recommended page with a redirect is an option. If one of your pages is dedicated to the topic but ranks worse, the task names it as a note.

With Choose another page you pick the page yourself. Your choice then also applies to the other tasks and to the link suggestions. Queries that several of your pages compete for only count towards the tasks on strong demand far down the results and on queries just behind page one for the recommended page.

Link suggestions only appear where a better position brings clicks

An internal link suggestion is only created if the target page ranks at position 4 to 20 for a query and the demand is reliable: at positions 4 to 10 it needs at least 50 impressions, at positions 11 to 20 at least 150. In addition, moving up five positions, at most to position 5, has to bring at least 3 more clicks a month.

If the target page already has the task on its queries just behind page one, rankmio holds the single link suggestion back. That group task already recommends internal links.

Tasks on title and description length do not count brand searches

Anyone searching for your own brand name clicks regardless of the title. Tasks on the length of title and description therefore do not count these brand searches. If fewer than 50 impressions remain without them, rankmio holds the task back and gives the reason.

Detailed feature description with screenshots and FAQ: View feature page

The audit compares any period

At the top of the drawer you pick two dates. Everything below then refers to that period. The audit is reached through the orange button in the "Top 5 Fixes" block or through Audit in the sidebar.

rankmio distinguishes two kinds of data, and the timeline below the date fields shows the difference. Metrics such as clicks, impressions and the visibility index are recorded daily, so every date has a value. Findings such as snapshots, GEO audits and citation scans only exist where something was measured. The timeline marks exactly those days, and clicking a mark sets the start date there.

You then see a balance in five tiles (visibility index, clicks, impressions, AI citations, completed tasks), the change against the equally long period before for clicks and impressions, and trends of your choice one below the other on a shared time axis.

Clicks and impressions are Google and Bing combined. The average position stays Google-only and is labelled as such, because Bing provides no comparable figure.

Where a source is missing, the audit says "not collected" and never shows a zero. The difference matters: Zero means "measured, result nil", not collected means "not measured". Tasks whose basis has been deleted, for example because an analysis expired, do not count as completed either.

Completed tasks appear on the trend

Whatever you tick off appears as a mark below every trend. Hovering names the date, the count and the categories. That lets you connect a movement in the numbers with the work behind it; you draw the connection yourself instead of an automation claiming it for you.

The "Resolved problems" list at the foot of the drawer is project-wide, not per user. In a team it therefore counts everybody's work.

An AI summary describes the chosen period

A button writes a summary of the chosen period. It draws on the balance, the trends, the confirmed completions and the most important open tasks.

You are charged once per period and data state, and opening the same summary again is free. A fresh measurement makes a new version possible. The price is on the price overview.

The summary must not invent numbers, must not guess where a source reports "not collected", and must not claim a cause the completed tasks do not support. If nothing was confirmed in the period, it says so plainly.

The PDF report inherits the chosen period along with the balance and the trends. What you assemble on screen is what your client receives, without a second round of settings.

The analysis status shows which analyses have run

The Analysis status button opens a panel that answers two different questions.

The ring shows the coverage: how many of the 18 analyses have already run for this project. Below it every analysis is listed with its date and state, and where possible you can start a missing one straight from there.

Below that, as a secondary figure, stands the rankmio Score: the weighted average across the graded pillars. Eleven of the 18 analyses carry a grade; the remaining seven are only counted, because for them there is no meaningful good or bad, only present or absent.

PillarWeight
Performance (PageSpeed)16%
SEO Checklist16%
Keywords12%
GEO score10%
Core Web Vitals8%
Sitemap8%
Battle8%
Content & AI6%
Content Optimizer6%
Citation Test6%
robots.txt4%

The colours mean: green from 70, amber from 40, red below 40, grey not analysed.

The visibility index adds up four sources

The visibility index is a value from 0 to 100 built from four sources with fixed shares: SERP visibility 40 per cent, AI citations 25, traffic 20 and the backlink profile 15. The total is their sum, not their average.

Missing sources cost their points and are not redistributed across the remaining ones. A project without a backlink analysis therefore reaches at most 85 points, and the panel states that ceiling; otherwise a capped value would be indistinguishable from a poor one.

The SERP pillar weighs demand and position

Everything rests on one quantity, visibility: the sum across all search terms of demand times the click probability of that position. A term in position 1 with 1,000 searches contributes 280, the same term in position 40 only 4. That is the difference from a pure keyword count, which treated both alike.

Two parts follow from it. The level (60 points) measures size on a logarithmic scale up to an anchor of 3,000 expected clicks. The capture rate (40 points) measures quality: how much you extract from your own demand compared with holding position 1 everywhere. Only queries up to position 20 count here; new queries further back raise the level and do not lower the capture rate. Below 300 impressions a month the capture rate is not calculated, because with four impressions it would inevitably read as one hundred per cent; those projects reach at most 60 of the 100 SERP points.

Three sources feed this pillar. The Search Console provides the basis and wins every overlap, because it measures actual impressions instead of estimating them. The position tracker only adds terms the Search Console does not know at all. Bing clicks are added on top, while the ranking part counts Google only.

The other three pillars use their own measurement

The AI citations pillar is the average across the checked LLMs. For ChatGPT, Claude, Gemini and Perplexity, rankmio takes the answers from the citation tracker and discovery of the last 28 days and calculates the citation rate times 80 plus the rate of "cited or mentioned" times 20. Counted are your active questions and the questions you checked in discovery; questions with your brand count the same as topic questions. An LLM only counts from ten answers. Google's AI features come in through the citation tracker. The uploaded reports from Bing and the Search Console are shown alongside and do not count towards the value, because they only report citations and no answers without one.

The two files count quantities: a page that is cited at all gets 30 points, and above that the value rises logarithmically, so every tenfold increase adds the same number of points. A file older than 30 days counts with its monthly average until day 90. Next to the pillar you see how many of the six LLMs are checked, because unchecked LLMs are not part of the average.

The traffic pillar puts Google and Bing clicks together on a logarithmic scale, where 10,000 clicks in 28 days give the full score. The backlink pillar measures the strength of the referring domains: how many there are and how strong they are. Links with a high spam score count neither for nor against you, and domains that Google or Bing also report as linking count more.

A new snapshot is written in the nightly run, which starts at 02:00 and calculates the index after the Search Console and Bing data have been collected. The history, the sparkline and the seven-day trend are built from these snapshots. A paid SERP measurement you trigger yourself stays valid for seven days and is carried through the nightly runs.

3GEO Module (AI Visibility)

The GEO module (Generative Engine Optimization) shows how well your website is prepared for AI answers and how often AI systems actually cite it. Its core is the GEO audit; around it sit the citation tracker, the reports from Bing and the Search Console, and the tools for individual pages.

The GEO score consists of citation readiness and actual citations

The GEO score from 0 to 100 is half citation readiness and half actual citations. Citation readiness counts which prerequisites your pages fulfil so that an AI can cite them. Actual citations measure whether your pages demonstrably appear in AI answers.

Citation readiness is measured on the HTML of each page, and every point is a rule you can check on the page yourself. Four points are mandatory, because Google and the AI providers name them themselves: search crawlers may fetch the page, it is indexable, snippets are allowed, and the content is available as text. If a page misses one of them, it cannot be cited, and its citation readiness counts as 0. Three structure points follow: one H1, a clean heading order, and lists or tables in longer texts. For articles, author and date are added. Two further points are notes that cost no points: paragraphs over 200 words, because Bing advises against walls of text without naming a number (rankmio sets the limit), and an AI's verdict on whether the opening states the core message of the heading. FAQ markup deliberately does not count, because pages with FAQ schema were cited less often in a large evaluation.

Actual citations come from three measurement sources per page: the citation tracker contributes up to 40 points, Google's AI report from the Search Console 30, and Copilot from the Bing Webmaster Tools 30. If a source is not connected, its points are missing, and the score says "of at most …" so that a capped value cannot be mistaken for a poor one. The citation tracker counts the answers of the last 28 days, the same as the citation rate. An uploaded report counts as long as its period ended no more than 60 days ago; after that, upload a new one.

The GEO audit evaluates up to 400 pages

You start the audit on the GEO page and choose how many pages it evaluates. rankmio recommends a number based on the size of your sitemap. The audit runs its own sitemap crawl and then evaluates every page. The GEO page then shows the latest audit. At the top are four figures: the GEO score, the citation rate (the same as in the citation tracker), blocked pages and orphan pages. Below them, a list names the five biggest levers, the biggest first, with the AI summary to expand. A single trend shows the GEO score per audit and the citation rate per scan, and switches to the citation rate per AI system. The details are spread over seven tabs.

TabWhat it shows
RequirementsPages that miss at least one requirement, mandatory points first; below them structured data and the access of the AI crawlers
Orphan pagesPages that no other page links to
All pagesThe evaluation by page type, then every evaluated page with its score; per page the way into the Content Studio and the schema generator, which creates the matching JSON-LD markup to copy
LinkingLink recommendations first: which page should link to which, with the reason and a button that opens the source page in the Content Studio. Each recommendation also appears as a task in the task list. Below that come the link distribution, orphan pages and all links.
AI search phrasesA pointer to the Discovery of the citation tracker, where the questions for AI visibility are created now; existing prompts can be imported from here
Gap analysisThe entity gap analysis and the content gap analysis
Previous auditsEarlier audits of the project with their results

The audit also points out where the bottleneck lies. If the pages score high on citation readiness but are rarely cited, the problem lies outside your pages; links from other websites are most strongly related to citations. A further note checks whether the first paragraph of a page already delivers the core statement its heading promises. Because this assessment comes from a language model, it appears as a note and does not count towards the score.

The trend covers the last 90 days. Audits from before 5 October 2026 were calculated with the old formula and appear as a separate grey line.

Further tools sit inside the audit

Schema.org generator

Generates JSON-LD markup matching the page content, for example Article, Product or Organization, from the button "Schema" in the "All pages" table. It only uses values that appear on the page. You copy the code straight into your page; once created, it is stored and opens again free of charge.

Entity gap analysis

Compares the entities your website covers with up to three competitors and shows which topics and entities are missing.

Content gap analysis

Compares your domain with one to three competitor addresses you enter and shows which topics and entities they cover and you do not.

Three sources measure actual citations

The citation tracker actively asks AI systems your questions and records whether your domain is cited. Two reports below it show what real users have triggered. Bing AI Performance shows Copilot citations from a CSV export of the Bing Webmaster Tools. GSC AI Performance shows impressions in Google's AI answers from an export of the Search Console. Both reports feed the actual citations in the GEO score and count as Copilot and Google AI among the checked LLMs in the visibility index.

How often your site is cited in the citation tracker depends heavily on which questions you have checked. The most meaningful questions are the ones your readers would really ask. Questions with your name or domain count the same as all others, because you asked them deliberately. Below the citation rate you therefore see how you do for questions with your brand and for topic questions, each with an assessment: for brand questions a citation is the expectation, for topic questions 20 per cent is already a good result. Real questions from Copilot users in the Bing report can be taken into discovery free of charge and measured there.

AI Champions lists pages that AI systems cite but that do not (yet) rank on Google.

The citation readiness check evaluates a single page

Under Page Analysis you look at a single page. For free, rankmio shows what it already knows about the page: the most important search queries of the last 28 days from Search Console with position, impressions and clicks, and the citation readiness from the latest GEO audit with its date. If the page does not rank in the top 20 for any query, rankmio points out that a better title does nothing there, because the page first has to climb in the rankings. A new measurement costs credits. It shows the SEO structure and the citation readiness with the date and lists the open points that follow directly from the measured values, such as a title that is too long, missing alt texts or an unmet citation readiness point. With «Revise in the editor» you open the page in the Content Studio, and the open points come along as a note. The editor keeps measuring the same points for free while you write.

The citation test simulates an AI answer

The citation test asks an AI web search a question, checks whether your page is cited, and measures how well the page fits the question and how citable it is. Details are under 7. Tools → citation test.

Detailed feature description: GEO/AI Hub feature page

4SEO & Technical Website Analysis

The website analysis is the heart of rankmio. It checks technical SEO factors, PageSpeed, Core Web Vitals, crawler data, and more — for your homepage and subpages.

You analyse a page in three steps

1
Go to Analyze → SEO Analysis and enter the URL, or pick it with the URL Picker.
2
Click Start Analysis. The results are ready after about 30 seconds.
3
If you like, have AI action items generated, which are prioritised measures with an impact rating. They come from queries just outside the top 10, from ranking losses and from seasonal peaks. A ranking loss only counts over the queries that occur in both periods, so that new queries far down do not fake a loss. Pages that rose create no task, and cannibalisation as well as title and description are judged by their own checks.

The URL picker offers pages from three sources

The URL Picker is a unified page selection dialog available on many pages (Analysis, Page Analysis, Citation Test, GEO, Competitors). It offers three sources:

Sitemap URLs

All URLs from the latest sitemap crawl of your project. Requires a completed sitemap crawl.

Project Pages

URLs already saved in your project (e.g. from previous analyses or subpage batches).

GSC Pages

Top pages from Google Search Console (sorted by clicks). Requires an active GSC connection.

The URL Picker features a free-text search and pagination (20 entries per page). Clicking a URL inserts it into the input field.

The results cover four areas

PageSpeed

Performance, Accessibility, Best Practices, SEO — mobile + desktop. Including Core Web Vitals (LCP, INP, CLS) with traffic light rating.

SEO Crawler

H1-H6 heading hierarchy, meta tags, images (alt texts), internal/external links, Schema.org, canonical, HTTP status, redirect chains, robots.txt check.

Visibility (analysis score)

A value from 0 to 100 built purely from the results of this analysis: SEO score 30 per cent, on-page checks 25, performance 20, accessibility 15, best practices 10. No keyword data goes into it. Not the same thing as the visibility index on the dashboard, which is built from four entirely different sources.

SEO Health Checklist

Extended technical check with DOM analysis. Automatic checklist with concrete improvement suggestions and per-URL todos in the site audit.

The subpage analysis checks up to five pages at once

Checks up to 5 subpages per batch (max. 20 per project) with the same depth as the main analysis. Select URLs via URL Picker (GSC, Sitemap, Project Pages) or enter them manually. Optional AI audit per page.

1
Go to Analyze → SEO Analysis and click Subpages.
2
Select up to 5 URLs via URL Picker or manual entry.
3
Start the batch — each URL is analyzed individually.
4
Optional: AI audit per subpage for prioritized recommendations.

The analysis history shows the development

A chart shows the development of all metrics over time. Shows trends and the impact of your optimizations. CSV export for all history data is free.

Detailed feature description with screenshots and FAQ: View feature page

5Keyword-Hub

The Keyword Hub is the central keyword cockpit. It brings all keyword functions together in twelve tabs, and on the left a page list lets you narrow everything to a single page. Each project can store up to 500 keywords; pagination, filters and sorting keep even large lists manageable. Search data from the Search Console covers the last 90 days.

The research tab finds new keywords

The Research tab bundles keyword discovery in three sections. Under My Domain rankmio finds the terms your domain already ranks for, without you entering anything: the quick scan returns 50 keywords, the bundle 200 with clusters and an action plan, the XL bundle 1,000. Under Competitor Analysis the keyword gap shows terms a competitor ranks for and you do not. Under Manual Research the keyword finder combines Google Suggest, DataForSEO and the keyword density of a page, and the AI suggests terms from the page content. Costs: see Credit prices.

The page list narrows every tab to one page

The list on the left shows the pages of your project with search and pagination, a filter for pages with and without issues, and coloured markers for the type of issue. If you select a page, a banner shows the active filter, and every tab only shows data for that page until you reset it.

The hub opens with Opportunities

The keyword hub opens with the Opportunities tab on the far left. If you select a page, Opportunities shows only the opportunities of that page. Since 8 November 2026 there is no separate Overview tab any more; it only repeated extracts from Opportunities and Cannibalisation, and the key figures of a page are in the page analysis.

My Keywords holds your keyword list

Keywords enter your project in four ways. You type them in yourself, with Google Suggest helping. You import the top 20 terms by impressions from the Search Console with one click, free of charge. You let the AI suggest terms, questions and synonyms from the content of a page selected on the left. Or you adopt results from the keyword finder individually or in bulk.

The table is sorted by impressions and shows source, impressions, position (last 28 days) and the trend against the 28 days before for each keyword. After enrichment it adds search volume (SV) and keyword difficulty (KD). The switch More columns adds the measured SERP position, relative difficulty, the Google Ads click price (CPC, in US$ as DataForSEO supplies it), traffic value, intent, competition, the elements of the results page and the simulation. All columns can be sorted and filtered.

Clean up selects all keywords from the Search Console import that had no impression for 90 days. You untick the ones you want to keep and delete the rest with one click. Keywords you added yourself or took from research are never suggested. The live position check asks for confirmation with its price first, because it costs a multiple of the standard check.

The Enrich SV + KD button fetches these market data from DataForSEO. Only keywords without a valid cache entry are queried, and a preview shows costs and cache status beforehand. The cache is valid for 30 days, so enriching the same keywords again within that time is free.

Opportunities shows where work pays off

The Opportunities tab calculates the best chances from the Search Console data, free of charge. The click gain shown is a potential with a named target position and not a forecast. The list is sorted by score, and the query you came from in the dashboard is at the top and highlighted. Queries with too little data are hidden by default; a switch shows them.

The reliability on each opportunity tells you how dependable the estimated click gain is. It rises with the number of impressions, measured against the minimum for the position, and falls if the position fluctuated strongly over the last 28 days. From 60 % the estimate is reliable. It does not measure whether the better position is reachable; that is what the difficulty shows.

Every opportunity has the button Implement in Content Studio. It opens the page in the editor with the query, position, target and click gain as the task. If the term is already in the title and H1 of a page with enough text, the task recommends internal links and links from other websites instead of more text. For snippet cases it names the actual and expected click rate, and if several of your pages rank for the query, it asks you to decide on one page first.

An opportunity marked Win back concerns a search query on page 1 or 2 that lost positions over the last 28 days. rankmio states where it ranked before and a possible cause: another of your pages now ranks for it, the page was changed during the period, or no change to the page was measured. The Win back filter shows only these opportunities.

rankmio calculates no gain for queries that bring not a single click on page one although chance can be ruled out, or whose demand is not confirmed: more than half the impressions on three days at most, a question from Google's AI Mode or no search volume. They are listed with the reason in a collapsed block below the list and are not counted in quick wins, missed clicks or the biggest opportunity. The Coverage tile says how many of the clicks of the last 28 days Google assigns to a query at all, because only this part can show up in the list.

Cannibalization finds competing pages

The Cannibalization tab shows search terms for which several of your own pages compete in Google, with all ranking addresses and the distribution of impressions, sorted by impressions. It is free. Old addresses that redirect to a page, and the same address with and without www, count as one page; the entry says which addresses were combined. Language versions such as /fr/ or /en/ serve other markets and are not counted as competition, nor are questions from Google's AI Mode. Several of your own pages for your own name are normal as long as one of them is in first place; such entries are at the end.

Every page of a conflict has the button Implement in Content Studio with its own task: the strongest page is to become the clear main page, the weaker ones set themselves apart or are merged into it. If no page is clearly ahead, the task asks you to decide which page should serve the query.

In the task list, rankmio combines competing pages into one task per topic with a recommended page. How the recommendation is made is described under Cannibalisation in the tasks.

A notice shows when stray traffic skews your position

If at least 5 % of your impressions come from countries outside your main market and this shifts the average position by at least half a place, the Keyword Hub shows the notice “Stray traffic is skewing your average position”. It names the share, the position shown and the position for your main market alone. rankmio derives the main market from the impressions. The stored figures stay unchanged so that the history shows no artificial jump.

The action plan turns the data into measures per page

The Action plan tab gathers the issues of your pages and lets an AI write concrete measures from them. Three kinds of data feed in: queries close to the top 3 in positions 4 to 10, ranking losses per page and seasonal peaks. A ranking loss only counts if a statistical test confirms it over 28 against 28 days and the page was in the first 20 places before or after; the position is weighted with the impressions of the earlier period, so a sudden flood of impressions far down does not distort it. Queries without confirmed demand or without a single click on page one do not feed in.

In the scope dropdown you choose how many pages the AI should process, in steps of 15. The result is a prioritised list with a reason, the affected keyword and a concrete instruction for each task. Measures are saved for seven days; a task whose opportunity no longer exists is hidden. Every task, free or written by the AI, has the button Implement in Content Studio, which opens the page in the editor with the task. Prices: see Credit prices.

Optimization groups keywords and analyses search results

The Optimization tab contains two tools. The SERP cluster groups your keywords by the similarity of their search results: terms that share the same results can be covered on one page. The SERP landscape shows the search results for a keyword with all elements such as featured snippets, "People also ask" and videos; from there you can mark competitors for a gap analysis or create a content brief. A link leads to the Snippets & click rate page.

The monitor watches a competitor domain

In the Monitor tab rankmio fetches up to 500 rankings of a competitor domain and shows with every run which search terms it has newly won or lost.

The Bing tab shows your search performance at Bing

The Bing tab shows clicks, impressions, click rate, average position, the most frequent search queries and the strongest pages from Microsoft Bing. The data comes from the Bing Webmaster Tools, either as a CSV export of the search performance report or via the connection; rankmio recognises the type of file itself. The prerequisite is that a Bing Webmaster Tools website is assigned to the project. Which of your pages Bing knows is shown in the sitemap area under Indexing.

Pages is a work list across all pages

The Pages tab answers one question: Which of your pages can you work on next with a prospect of measurable clicks? It brings the page directory and the search data together and sorts by click deficit, that is the clicks the best keyword of a page lacks compared with the expected click rate of its position.

Five filters narrow the list. Click deficit shows pages where a keyword demonstrably falls short of the expected click rate. Lever shows pages with at least one keyword between positions 4 and 20. Visibility falling shows pages whose impressions dropped by more than 40 per cent in the last 15 days; there the cause has to be found first. With findings shows pages with open tasks, Without search data pages without a single impression in 28 days.

For each page the profile shows how many keywords stand in the top 3, in positions 4 to 10 and in 11 to 20. You can download the filtered list as a CSV table or as a text file with one address per line.

Performance links keywords and load time

The Performance tab shows the PageSpeed scores for mobile and desktop, the Core Web Vitals (LCP, INP, CLS) and the performance history of the project.

Below it is the list of search queries on pages 1 and 2 that have moved. The last 28 days are compared with the 28 days before, and for each query the best own result counts. Only queries with enough impressions for a reliable verdict are included: from 20 at positions 1 to 3, from 50 at positions 4 to 10, from 150 at positions 11 to 20. The list is sorted by the clicks a movement is estimated to gain or cost in 28 days. Winners lead to the page analysis, losers straight into the editor.

History shows positions over 90 days

The History tab shows the positions of your keywords over the last 90 days as a chart. With the page list on the left you narrow it to a single page. It is free.

Alerts report position changes

In the Alerts tab you watch individual keywords. You choose a keyword, a threshold such as position 10 and whether the alert should fire when the keyword rises above or falls below it. rankmio checks the alert with the next Search Console sync; a triggered alert appears as a banner on the dashboard, in the alert history and, if you like, by email. It is free.

Snippets & click rate shows where a better title pays off

Between a good position and an actual click sit two lines of text: the title and the description Google shows in its results. The page under /snippets/{project} answers where changing them pays off and gives a reason for every answer.

Each query receives exactly one verdict: snippet, position too far back, two own pages, environment in the result list, demand falling away, different search intent, or too little data. Only the first leads to a proposal, because a page on position 47 has a ranking question and not a title question, and a page whose impressions are collapsing needs the cause found first. Searches for your own name and questions addressed to an AI are listed separately and stay out of the assessment; queries without anything conspicuous are marked as such.

A verdict is only given from a minimum number of impressions in the 28 days of the measurement period: from 20 at positions 1 to 3, from 50 at positions 4 to 10 and from 150 at positions 11 to 20. Below that, the verdict is “too little data”.

The findings are arranged in three areas, ordered by the moment at which they take effect.

1
In the first area a better snippet wins clicks right away. These pages rank well, their click rate still falls reliably short of expectation, and no other cause explains it.
2
The second area gathers what works once the page climbs. Two figures stand side by side: what the snippet brings today, and what it would bring at a better position. If you intend to move one of these pages forward anyway, write the title along the way.
3
The third area holds pages that lack a description of their own. Where you have stored none, Google assembles a snippet from your page text. This is worth doing regardless of position.

rankmio derives your brand terms from the domain name, suggests further ones, and you confirm them. You can also pick any page of your project yourself, including one with no impressions at all. The assessment costs you nothing; only the generated proposal costs credits.

The proposal grows out of the page's real search queries, its search intent and the device mix, which is why a project with a mobile majority receives a shorter title. Title and description sit in input fields: what you apply is your own wording, and it is stored separately from the proposal. What a change achieved is then compared before and after, though only at a comparable position. Prices: see Credit prices.

If you changed the title or description yourself, without a proposal from rankmio, add the change under “Add your own change”. rankmio records the state of the page on that day as the before value and measures again after two weeks, just as with an applied proposal. This costs nothing.

Every page has one card: the strongest query leads, the further queries of the page are listed below, and there is one button, because a title applies to the whole page. Implement in Content Studio opens the page in the editor with the diagnosis as the task; after Generate proposal a second button hands over the proposed title and description, including your edits. Beyond position 20 the proposal button is shown in a muted style with a note on why it does not pay off yet. Queries that bring not a single click on page one count as a different search intent, and queries whose demand is not confirmed are marked as such. Files such as .txt or .pdf do not appear under pages without a description. If title and description of pages are still unknown, a note at the top links to the sitemap crawl.

Confirm your brand terms first. Searches for your own name are clicked almost regardless of the wording, and keeping them out sharpens every figure that follows.
Detailed feature description with screenshots and FAQ: View feature page

6Competitors

The competitor analysis offers a central register of pinned competitors, loadable data packages per competitor, three comparison modes and an automatic battle schedule. All results refer to the active project. The sitemaps of competitors are monitored in the Sitemap Hub under Sitemap Monitor, together with your own.

The register holds your pinned competitors

The foundation for all other competitor functions. Under Competitors → Register, you manage the most important competitor domains of each project. Pinned competitors are used automatically in other modules such as backlinks, position tracker, GEO and citation tracker, and the chat agent keeps their brand terms out of its keyword recommendations. For briefs in the Content Studio, at most three pinned domains are used.

1
Add a domain (manually or via Competitor Discovery from SERP and briefing data).
2
Pin with the star icon — only pinned competitors flow into other modules.
3
Optional: Load data package (see below) for rankings, backlinks, historical overview.

A data package loads everything about one competitor

For each pinned competitor, you can load a complete data package with one click — rankings, top pages, a backlink summary and a historical overview. What it costs depends on the components you choose; the prices are on the price overview.

  • Rankings — up to 1,000 ranking keywords of the competitor (position, search volume, KD, intent, ETV) — the data basis for gap analyses and keyword takeover.
  • Top Pages — which competitor URLs draw the most traffic, sortable.
  • Backlinks Summary — referring domains, anchor distribution and spam scores.
  • Historical Overview — how visibility has developed over the past months.
Tip: Data packages are cached. Repeated calls cost no extra credits; refresh button reloads current data on demand.

The quick compare checks the home pages of up to three domains

Compares the home page of your site with the home pages of up to 2 competitors. You find it under Website comparison via the switch at the top.

1
Go to Competitors → Website comparison, switch to the quick compare at the top and enter up to 2 competitor URLs.
2
Click Start Compare.
3
Result: comparison matrix with PageSpeed for mobile and desktop, load times, a technical score and basics such as H1, canonical and schema. The technologies in use (CMS, frameworks, analytics, CDN) are only detected by the deep analysis.

The deep analysis crawls a competitor in depth

Crawls 50 or 100 subpages of a competitor and delivers a detailed comparison analysis with AI insights.

  • Content analysis: heading structure, text length, internal linking
  • Technology stack: CMS, frameworks, CDN, analytics
  • AI insights: Prioritized action items based on the comparison
  • PageSpeed of the home pages for mobile and desktop with load times
  • Export as standard or branded PDF

The battle compares two pages one to one

The battle compares one of your pages with a competitor page in two areas. Under SEO metrics it sets on-page score, load time, meta description, headings, word count, images, links and canonical side by side. Under GEO / Citation Readiness it shows the measured AI citation readiness of both pages, whether each is cited in the same AI answer, the search result position, freshness, schema and author. Where a value could not be measured, a dash stands in its place instead of an estimate. A colour-coded matrix and a total score show who leads where. The total score combines citation readiness (70 %) and the on-page score (30 %).

Tasks from the battle follow the most recent comparison. Once you have changed the page, older recommendations expire.

A schedule repeats battles automatically

Automate recurring battles with the Battle Scheduler:

1
Define the competitor domain and your keyword/focus.
2
Choose the frequency: daily, weekly, or monthly.
3
rankmio runs scheduled battles on its own and checks every 30 minutes whether one is due. You find the results in the battle history.

Schedules can be paused or deleted at any time. Credits are deducted with each execution.

The chat agent filters competitor keywords

The chat agent uses your pinned competitors automatically when it answers keyword questions:

  • Brand filter — keyword recommendations never include the competitors' own brand terms. You only see keywords you can realistically rank for.
  • Topic filter — competitor keywords that are topically off-target for your site (compared to your last 90 days of Google Search Console data) are dropped before the agent sees them.
  • Transparent — when you ask the agent why a keyword was excluded, it can show you exactly which filter removed it (brand match, off-topic, etc.).
Detailed feature description with screenshots and FAQ: View feature page

7Tools

The Sitemap Hub crawls your website and checks the index

The Sitemap Hub has three tabs at the top: Sitemap crawls, Sitemap monitor and Indexing.

A sitemap crawl starts at the address of your project, which you can change before starting, and follows every internal link. A progress bar shows how far it has got; a crawl that hangs is stopped after 15 minutes and the credits are refunded. Pages that could not be loaded while links were being discovered are named after the run, because links reachable only through them are missing from that crawl.

The results are spread over eight tabs.

TabWhat it shows
OverviewKey figures of the crawl
SEO HealthTechnical findings per page, such as titles that are too long or missing headings
Dead linksPages with HTTP errors, each with the page the link stands on, and the pages that did not answer at all
Sitemap gapsReachable pages with sound content that are missing from your sitemap.xml
Thin contentPages in your sitemap with fewer than 500 words whose content was rated weak
Older datePages whose content has changed since the last crawl while the date in the sitemap stayed older
All URLsEvery address found
Sitemap editorAn optimised sitemap to adjust and download as XML

If your existing sitemap.xml could not be read during the crawl, no comparison takes place. rankmio then reports no gaps and says so on the results page, instead of listing every page it found as missing. The "older date" tab stays empty until the second crawl, because the comparison needs an earlier run.

The Sitemap monitor watches your own sitemap and those of competitors and reports new and removed addresses.

The Indexing tab shows whether Google and Bing know each page

For every page of your last sitemap crawl, the Indexing tab shows whether Google has taken it into its index and, if not, why, and whether Bing knows it. The Google status comes from the URL inspection of the Search Console, the Bing status from the Bing Webmaster Tools. The prerequisite is that the project is connected to the Search Console or has a Bing Webmaster Tools website assigned.

For Google, rankmio distinguishes indexed, blocked or fetch error, read but not taken in, another page authoritative, known but not fetched, unknown to Google, and a few rarer states. Next to each status stands Google's own wording, the date of the last fetch and, where Google regards another page as authoritative, its address. Pages that rankmio's own crawl already knows as redirects, as noindex or with a foreign canonical address are not queried at Google; there are separate tasks for them.

For Bing, the tab only shows whether Bing knows a page and when it last fetched it. Bing reports through no interface whether a page is in its index; for a single page, the URL inspection in the Bing Webmaster Tools shows that. You can report pages Bing does not know to Bing individually or all at once. Filters show only the pages Google has not indexed or Bing does not know.

Google reports the index status with a delay of several days. rankmio checks new pages in the following night and known pages again once a week; you can also update the status by button. A Bing status is valid for seven days. Pages that Google does not index and pages Bing does not know become tasks in the task list, with an explanation of what you can do.

The robots.txt tool edits and checks your robots.txt

The robots.txt Tool offers seven functions for managing and analyzing your robots.txt:

  • Editor — Edit robots.txt with automatic version history
  • AI Generator — Let AI create an optimal robots.txt
  • AI Analysis — Check for errors and optimization potential
  • Live Check — Fetch current robots.txt from a domain
  • CSV Bulk Test — Test multiple URLs against rules
  • Monitoring — Automatic monitoring with email alerts on changes (every 30 min.)
  • Versioning — Compare all versions (diff view). Free.

Costs per function: see Credit prices.

The parser follows the Google specification: wildcards, $-anchors, longest-match-wins. It knows 22 bots in four categories: 8 search engines, 9 AI crawlers, 3 social networks and 2 others.

The link profile check builds a disavow file on request

The Link Profile Check shows you which patterns rankmio has observed in your backlinks, and builds a disavow file for Google Search Console from them at your request. It requires backlinks to be loaded for the project in the Backlink Hub ; without that data the page stays empty. Besides the data from DataForSEO, you can upload the link exports of the Search Console (ZIP or CSV) and of the Bing Webmaster Tools in the Backlink Hub; they extend the coverage. The function costs no credits.

rankmio does not classify any link as toxic and makes no recommendation. The application shows a pattern and names the evidence for it; the decision is yours. A pattern is an indication, not proof: a naturally grown link profile can show one of these characteristics too.

Links are grouped by these six rules:

  • Anchor text — the same anchor text appears on at least three external domains. Anchors that merely name your own domain are left out, because they occur with every mention.
  • Address pattern — at least three external domains link through the same path structure, for example always from the same directory.
  • Clustering in time — at least five referring domains were first seen within no more than seven days.
  • Uniform addresses — a single domain links from at least five pages whose addresses differ only in a sequence of digits.
  • Your own projects — the links come from your other projects in rankmio. They are marked as harmless and cannot be added to the file.
  • Follow status — the links carry a conspicuous spam score but are marked nofollow, so under the usual handling they pass on no ranking weight.

Every group carries one of three levels: conspicuous, indication or harmless.

Three checks come before the file. The download button only becomes active once you have confirmed all three:

1
You have checked in Search Console whether a manual action for unnatural links exists.
2
You have checked whether a disavow file is already stored for this property. A new upload replaces the existing one entirely.
3
You have compared the marked domains against your own traffic data.

For each group you choose between «Do not include», «Include whole domains» and «Include individual addresses only». Covering a whole domain includes every link from it, an editorially placed one too. In the free-text field you can add your own entries, one per line; a line beginning with a hash becomes the reason for all entries below it.

You upload the finished file to Google Search Console under «Disavow links». It applies per property: if you have set up your website both as a domain property and as a URL prefix, store the file with the property under which the links are reported.

In depth, with examples and the limits of the method: our article on the disavow file

Page Analysis measures an existing page and leads into the editor

Page Analysis (formerly Content Optimizer and Citability) is a measurement sheet for an existing page. The open points come from the measured values, not from free AI text, so they do not contradict the figures next to them. You revise the page in the Content Studio editor, which measures title, description, headings and citation readiness live. Every measurement is stored in the history with its date and can be loaded or deleted there; if it is older than 30 days, the page says so.

The citation test simulates an AI answer

The citation test is an AI answer simulation: it shows how AI search engines would use your page as a source. URL selection via the URL Picker.

1
Enter a URL (via picker or manually), optionally a question and city for local relevance.
2
rankmio asks the question in a real web search via ChatGPT and checks whether this exact page or another page of your domain is cited.
3
The score is measured, not estimated by an AI: fit to the question (40 %) and citation readiness as in the GEO audit and the page analysis (60 %). If the fit is below 50, the score stays at 40 at most. The actual citation is shown next to it and does not count towards the score, because a single answer varies.
4
A list compares the AI answer with your page: which points your page covers, and which points the cited sources mention and your page does not, each with its source. The top 3 actions follow from the weakest values and these gaps.

You can check saved simulations again; a badge shows the change, for example +12 in green or -5 in red. Up to 20 results per project are kept.

Besides the standard run, Multi asks the same question three times and Premium five times. The measured values do not change between runs; what you see is in how many runs the AI web search cites the page.

The Open Industry Ranking on the same page shows which brands are most visible in AI answers on a topic, colour-coded by rank and with recommendations for action.

Free alternative — Chat Agent: Click "Ask Chat Agent" on the citation test page to open the Content Studio. The Chat Agent generates 5 AI questions with simulated answers — completely free. Each question+answer appears as an insertable card that can be added to your content with one click. The Chat Agent does not perform a real web search but provides a quick assessment. If a keyword is already entered, it is automatically included in the prompt.
Detailed feature description with screenshots and FAQ: View feature page

Some tools work without a login

These run on rankmio.de/seo-tools, need no account and cost no credits. One of them is limited because it triggers a paid model call.

1
Snippet check — measures title width in pixels, compares it against other pages on the same site and proposes a better version. One check per day, because a model call and a live search query sit behind it.
2
robots.txt test — reads the file and shows which search and AI crawlers you allow and block.
3
Sitemap tester — counts URLs, finds index files, reports XML errors and checks a sample of the URLs.
4
Schema.org check — extracts the structured data of a page and validates it.
5
Speed test — loading times and Core Web Vitals via the Google PageSpeed API.
6
Domain check — DNS records and WHOIS data for any domain.
7
Word counter — words, characters and reading time for any text.
8
My IP and password generator — two small helpers that run entirely in the browser.

8Reports & Export

rankmio offers various export formats for reports and analyses — as standard PDF (free), branded PDF (with your own logo), or CSV for further processing.

Most analyses can be exported as PDF

SEO Analysis PDF

All results as a printable overview. Free.

SEO Analysis Branded

With your own logo and company name.

Site Audit PDF

Action items with priorities. Free.

Site Audit Branded

Site audit with your own branding.

Subpage Analysis PDF

Results of the subpage analysis. Standard (free) or branded.

Competitor Deep Analysis PDF

Detailed comparison report. Standard (free) or branded.

Only the sections you have expanded on screen end up in the PDF; collapsed sections are left out. That way you decide what your client sees.

The analysis history (visibility, PageSpeed, Core Web Vitals) can be downloaded as an Excel-compatible CSV file free of charge.

The JSON export hands all project data to an AI coding agent

In the project header of the dashboard and the analysis pages, the JSON export button opens a dialog in which you decide for each data block whether it goes into the download, with the number of records and the estimated size. The blocks cover project data, on-page audit and crawler, PageSpeed, visibility index, GEO audit, citation tracker, backlinks, competitors, keywords, sitemap, the task list and the Bing search performance. A typical export is 200 to 500 KB of structured JSON that AI coding agents such as Claude Code, Cursor or GitHub Copilot can read directly. The task list in the export is the same as on the dashboard. Prices: see Credit prices.

The JSON API delivers the same data by request

The same blocks are available through an HTTP API, for pipelines and agents that fetch the data themselves. For the data export, three endpoints sit under /api/v1/: GET /projects lists your projects, GET /projects/{id} returns the project data and which blocks are available, both free; GET /projects/{id}/export?blocks=… returns the same JSON as the download. It costs credits when new measurements have arrived since your last paid export. If it returns exactly the same data as an export you have already paid for, it is free, and the same applies to the download. The response states in meta.charged_credits what was charged.

You authenticate with Authorization: Bearer rmk_…. You create, rotate and revoke keys under Profile → API keys; rankmio stores only a hash, and the key itself is shown once. Each key may make 60 requests per minute and 1,000 per day and spend at most 200 credits per day by default. Every call is logged. The error codes are 401 (not authorised), 402 (not enough credits), 403 (project or IP not allowed), 404 (project not accessible) and 429 (rate or credit limit reached).

The API offers the agent tools as well

Every tool an AI agent can call via MCP is also available as a plain HTTP request, apart from the agent guidance (get_next_steps, manual search), editor links, fetching third-party pages, team management and publishing pages of the rankmio website. GET /api/v1/tools lists the tools available to your key, each with its input schema. GET or POST /api/v1/projects/{id}/tools/{name} calls a tool for a project, for example /api/v1/projects/12/tools/list_tasks; account tools such as get_prices run under /api/v1/tools/{name}. With GET the arguments go in the query string, with POST as a JSON object in the body. The answer contains the data under data and, if credits were charged, a note under note. Rights, packages, prices, the read fee and the limits are the same as for MCP; a tool that costs credits asks for confirmation in the same way.

AI agents such as Claude work with rankmio via MCP

AI agents such as Claude Code or Cursor can connect directly to rankmio through the Model Context Protocol (MCP). The agent then asks rankmio itself, for example which pages lose clicks, where your best opportunities are or which AI systems cite you, and works with the answers. It uses the same data as the chat assistant. In the Content Studio it can also create briefs, write drafts and send them to WordPress, if the API key allows it. It loads existing WordPress posts as drafts, revises them and writes them back into the same post. On the SEO side it starts GEO audits, citation scans, page analyses and crawls of the website, fetches fresh backlink data, ticks off or hides tasks, saves keywords, fetches new data from Search Console and reports changed pages to Bing via IndexNow. An agent starts a crawl at most every 30 minutes per project and fetches backlink data at most every seven days, because a crawl puts load on the website and backlinks change slowly. Google has no interface for this; there, URL inspection in Search Console remains.

An agent can also set up the position tracker and the Citation Tracker. rankmio then measures, even without Search Console, where your website ranks on Google and whether AI assistants cite it. The agent lets the question search suggest the questions for the Citation Tracker. Because a tracker charges every run, the agent first tells you the cost per run and per month, and rankmio only sets the tracker up after this confirmation. The tracker then shows that an AI agent set it up. The agent asks what makes sense next with get_next_steps: rankmio names missing and outdated measurements and the open tasks, each with its reason, tool and price.

When you create a key, you choose what the agent may do. “Read only” shows data. “Read and act” also lets it create briefs, tick off or hide tasks, save keywords, sync Search Console, report pages to Bing and send finished drafts to WordPress as well as load, revise and write back existing posts; it asks you before every transfer and shows you what changes before writing back. “Read, act and spend credits” also lets it generate drafts and start research, GEO audits and citation scans. What your role does not allow in the web interface, the agent may not do either. Prices are the same as in the web interface. Before anything that costs credits, the agent first names the price and only books after you have confirmed. Each key spends at most 200 credits per day. Deleting is not possible via the agent. A new password does not lock API keys; you revoke a key individually in the API keys tab.

You connect with an API key from Profile → API keys. In Claude Code, one command is enough:

claude mcp add --transport http rankmio https://rankmio.de/mcp --header "Authorization: Bearer rmk_…"

Other programs need the address https://rankmio.de/mcp and the header Authorization: Bearer rmk_…. The agent first calls list_projects and then passes the project id to every other tool. With get_prices it can name all prices in advance, with get_credit_history it shows the latest bookings of your account. For the traffic trend, click losses and citations it can choose a period itself, and get_period_report returns the same period balance as the dashboard. Reading measurement data costs credits once per project as soon as new measurements exist, such as a new day from Search Console, an audit or a crawl; after that every further query about this project is free until the next measurement (see Credit prices). The project list, prices, team and Content Studio content never cost anything. The limits of the API key apply: 60 requests per minute and 1,000 per day. If you restrict a key to certain projects, the agent only sees those. It only gets the tools that belong to your booked packages.

Branded PDFs carry your own logo

Branded PDFs carry your logo, company name, and contact details instead of rankmio branding.

2
Upload your logo (PNG/JPG, max. 2 MB, min. 200x60 px).
3
Enter company name, address, and footer text and save.
4
From now on, the "Branded PDF" button is available for all PDFs.
Tip: Ideal for agencies and consultants — deliver professional reports to clients with your own branding.

9Projects

rankmio knows two kinds of projects. A website project corresponds to a domain that rankmio measures and evaluates; up to 10 website projects are free, each further one costs credits. A content project has no website of its own and is used for writing in the Content Studio. All analyses, keywords and settings belong to one project.

You create a website project with its domain

1
Go to Projects, click New Project and choose Website project.
2
Enter the domain (e.g. example.com).
3
Optional: Assign the project to a group (e.g. "Clients" or "Own Sites"). Groups can be created, renamed, and deleted.
4
Add a comment to document the project's purpose.

A content project holds texts without a website of its own

A content project bundles drafts, briefs, personas, media and the library for a topic that has no website, or none that rankmio should measure, such as a client's blog or a series of articles. Content projects do not count towards the limit of website projects; each account can have up to 100.

1
Go to Projects, click New Project and choose Content project.
2
Enter a name and choose the language and the country for keyword research.
3
Optional: Describe what the project is about and name the industry. rankmio uses both for personas and texts.
4
Optional: Assign the project to one of your website projects.

Under a website project, a content project uses its website data

If a content project belongs to a website project, the editor sees the same data as the website project: The Search Console, the crawled pages for matching internal links and, if your account includes the SEO functions, the action items and link recommendations. The competitor register is shared as well; a competitor you pin in the content project also appears in the website project. If the website project has not crawled any pages yet, the editor says so and links to the sitemap crawl. A standalone content project has no website data, so the Search Console tab in the editor is missing.

You change the assignment in the project overview with Assign and Detach. Only someone who owns both projects can assign them. A content project never turns into a website project. If you delete a website project, its content projects remain with all their drafts and stand on their own afterwards.

In a content project, rankmio shows only what works without a website

While a content project is active, the sidebar shows the Content Studio, keyword research in the Keyword Hub and the competitor register. The dashboard, GEO, the analyses and all other areas that measure a website are hidden. If you open one of them anyway, you land in the Content Studio library. The chat assistant advises you on your texts instead of rankings and only offers tools that work without a website.

A content project takes its internal links from WordPress

Assign one of your WordPress connections to a content project under “WordPress and CMS”. The editor then suggests matching pages of your website for internal links, taken from your published posts and pages. Drafts, private and password-protected content, legal pages and, if Yoast is installed, pages set to noindex are left out. rankmio refreshes the list at most every six hours, and this costs no credits. Pages outside WordPress and redirects are not known here. Content projects are no longer crawled.

rankmio checks whether your website can be reached

When you create a project, rankmio fetches the homepage once, the same way its crawler does later. If the website does not answer at all, the project cannot be created. If a protection service such as Cloudflare or your hosting provider blocks the request, rankmio creates the project anyway and shows a notice: Search Console, Bing and the trackers work without the crawler, but audits and crawls do not. The notice explains how to allow the crawler; it identifies itself as RankmioBot. The same notice appears when an audit fails because of a block, or when more than half of the pages in a crawl were blocked.

The project settings connect Google and Bing

In the project you assign the matching Search Console property and the matching Bing Webmaster Tools website. rankmio suggests properties that fit the domain instead of assigning one silently. The index check, the Bing tab in the Keyword Hub and the Bing tasks require the Bing assignment.

Page roles keep unsuitable tasks away from special pages

Some pages have a purpose to which many recommendations do not fit: an imprint should neither rank nor be cited by an AI. Under Optimize → Page roles you mark pages as legal, function, tool or directory pages. rankmio keeps measuring them like any other page but drops the tasks that do not fit their purpose; errors such as an unreachable page remain visible. rankmio suggests roles for pages that look like it, and asks you to check a marking if a marked page receives search traffic.

The project dropdown switches the active project

Use the project dropdown in the top navigation to switch the active project at any time. All views (Dashboard, Analysis, Keyword Hub, etc.) update automatically.

Project pages only show results for the currently active project (property filter). Switch the project to see data from a different domain.

10Team

The Team feature enables collaboration on projects with shared credit usage.

You create a team and invite members

1
Create a team under Team — choose a team name.
2
Invite members by email address (recipient must be registered). The first two members are free; each further member costs a one-time fee, see the price overview.
3
Assign projects — Choose which projects each member can view and edit.

Two roles divide the rights

Lead (Creator)

Creates the team, manages members and project assignments, pays all credits, configures branding. All member actions are charged to the lead's balance.

Member

Works on assigned projects using the lead's credits. Cannot manage members or set own branding. Read-only branding from the lead is applied.

The team shares one credit pool

All member actions are charged to the lead's balance. Members' own credits remain untouched. The lead can see a usage overview per member in the "Credit Usage" tab.

Tip: Team members automatically use the lead's GSC connection — no separate connection needed.

Content projects are shared like website projects

You share each content project individually, just like a website project. When you create a content project under a website project, the box "Share with all team members who can see the website project" is ticked. In the sharing overview, the button + content projects next to a website project gives everyone who sees that website project its content projects as well. A project that a member creates belongs to the team, that is to the lead, and is shared with the member who created it. This applies to both kinds of projects.

Detailed feature description with screenshots and FAQ: View feature page

11Credits

rankmio uses a pay-per-use model without subscriptions. Each function has defined credit costs. Credits do not expire.

You buy credits as packages

1
2
Choose a package: Starter (100 Cr.), Pro (500 Cr.), or Premium (1,000 Cr.).
3
Pay by credit card or local payment method (via Paddle). Credits are credited immediately.
Credits do not expire, and there is no subscription and no automatic payment. Credits are deducted when you start an action or switch on an automatic check such as the citation tracker; their cost per run is shown under “Running checks”.

Team members use the lead's credits

If you are a team member, you automatically use your team lead's credits. Your own balance remains untouched. The lead can view usage in the Team section.

Running checks use credits only when they run

Some checks run automatically once you switch them on: the citation tracker, the position tracker, robots.txt and sitemap monitoring, scheduled competitor comparisons and domain monitoring. Credits are deducted only when a run takes place. Without a check switched on, your account does not use any credits by itself.

Under Running checks you see all of them in one place, each with its frequency, cost per run and expected credits per month. Above the list you see how many days your balance lasts at this rate. Every check can be switched off there; checks set up by the AI agent are marked as such.

There you can also set a monthly limit for automatic checks. Once it is reached, they pause until the start of the month and you get an email. Anything you start yourself does not count towards it. If your balance will last fewer than seven days, rankmio warns you once by email. If the balance runs out, nothing is charged into the negative: the checks pause, you get an email, and after topping up they resume at their next scheduled time. On the dashboard a line shows what was used automatically since the previous evening.

The price overview lists every cost

All feature prices — sorted by category and always up to date — on the Credits page.

12Profile

Under Profile you manage your personal settings and integrations.

You maintain your account data in the profile

  • Name and Email — Change your account details
  • Password — Set a new password (bcrypt-encrypted). All other sessions of your account end, and you receive an email about the change.
  • Sign out everywhere else — Ends every session of your account except the current one, for example on a computer where you stayed signed in.
  • Email Notifications — Configure rank alerts, battle results, sitemap changes

The industry sharpens the recommendations

Under Profile → Industry Topics you select your industry and adjust platform weighting. This affects visibility calculations and AI recommendations.

Your branding appears in the PDFs

Upload your logo, company name, and contact details for branded PDFs. See Reports → Whitelabel.

You choose which AI providers are active

rankmio supports 4 AI providers. You can enable/disable individual providers for your account:

ProviderUsage
Claude (Anthropic)AI analyses, content briefs, action items, GEO
ChatGPTCitation test with web search, content optimization, sitemap AI
Gemini (Google)AI citation readiness, topic analysis
PerplexitySource citation, fact checking

3-Tier System: A provider is only active if (1) the admin has enabled it globally, (2) the API key is configured, and (3) you have enabled it in your profile.

You connect the Google Search Console in four steps

1
Go to Connections → Google Search Console and click Connect with Google.
2
Select your Google account and tick the permission for the Search Console. Without that tick rankmio cannot read any data, and it tells you so right away.
3
rankmio suggests the properties that match your projects; you confirm the assignment.
4
The connection is active immediately, and the first synchronisation starts as soon as the property is saved. After that the data is synchronised every night in the nightly run starting at 02:00, with a second attempt at 05:00.
Team members automatically use the lead's GSC connection — no separate connection needed. The connection can be disconnected at any time.

You can switch or detach a property in the project

For projects with an assigned property, the project view offers two actions. Switch property opens the picker with all properties of the connected Google account. Detach removes the assignment; data already imported remains. If a property you have just created in the Search Console is missing, Reload properties fetches the list straight away; otherwise it appears within 24 hours. You find the button next to the picker and in your account under Google Search Console. The Bing assignment can be detached in the same way.

13Privacy & Security

Your account and project data are hosted in Germany. For AI features, content is transferred to AI providers, some of them in the USA; details are in the privacy policy.

You can exercise your GDPR rights yourself

  • Data Export (Art. 20) — Under Profile you can download all your data as a JSON file (data portability).
  • Delete Account (Art. 17) — In the Profile area you can permanently and irreversibly delete all account data, projects, analyses, and team memberships.
  • Right of Access (Art. 15) — The data export contains all information stored about you.

Several measures protect your data

  • Encryption — Passwords with bcrypt (cost=12), sensitive data with AES-256-CBC
  • HTTPS — All connections are SSL-encrypted (Let's Encrypt)
  • CSRF Protection — Every form is secured with tokens
  • Rate Limiting — Protection against brute-force attacks
  • Prepared Statements — Exclusively PDO Prepared Statements against SQL injection

This is how long rankmio keeps your data

Findings and analysis results are kept for 12 months, and the most recent result per project and type always remains. The daily snapshots of the visibility index are kept for 365 days, again with the most recent one always remaining. Search queries and daily values from the Search Console are deleted after 90 days, the index status of individual pages 365 days after its last check.

The cleanup is the first step of the nightly run, which starts at 02:00. Account data remains until you delete your account. The complete list of periods is in the privacy policy.

The servers are in Germany

Hetzner Cloud, Nuremberg (Germany). AI features use providers outside the EU, see the privacy policy. More details: Privacy Policy.

14SEO/GEO Chat Agent

The rankmio chat agent answers questions about your project with your own data. It reads the measurements rankmio has collected for the active project and explains what they mean. It does not start analyses itself; where a new measurement would help, it suggests the matching tool, and you start it yourself.

The agent works with the data of your project

Before every answer the agent takes in an overview of all measurement areas of the active project at once, from rankings, clicks and indexing to AI citations, backlinks and competitors, and it searches your project knowledge for passages that fit your question. Where data is missing, it says so and names the way to get it. The conversation is saved per project, so you can pick up where you left off. You can type your question or dictate it; the recording is transcribed and then discarded.

Where an analysis is missing or out of date, the agent shows a card with a link to the matching tool. A click takes you to the form with your question already filled in; the credits are only charged when you start the run there.

Diagnostic mode names three weak points

Questions like "Why doesn't my page rank?", "Why no clicks?" or "Why is my visibility low?" trigger the diagnostic mode. Instead of a data report, the agent names exactly three prioritised weak points. It draws them from seven areas: search intent, topic focus, page structure, citation readiness and E-E-A-T signals, technical SEO, internal linking, and the competition.

Each weak point is phrased by its effect first, for example "AI guesses your brand wrong" instead of "missing sameAs in schema". Afterwards the agent points out that there are further weak points, and a follow-up question such as "what else?" brings them up.

The agent can read these areas

The agent has more than fifty read-only tools, grouped into the following areas. None of them changes anything in your project or costs credits.

AreaWhat the agent can look up
Search ConsoleOpportunities, quick wins, keyword details, cannibalization, zero-click terms, page performance, traffic trend, positions by band, snippet analysis, click-rate gaps, lost clicks per page with their cause, and the monthly search demand of the last twelve months, from which the agent recognises seasonal keywords
Tasks and statusAction plan, project status, SEO health, sitemap findings, robots.txt status, Google's index status, visibility index month by month
GEOGEO values per page, citation readiness issues, answer quality, Google's AI report, Copilot citations, structured data per page, AI coverage gaps
CitationsCitation overview, cited phrases, competitors in AI answers, newly discovered citations
WebsitePage content, internal links, broken links, PageSpeed results, duplicate content
CompetitorsPinned competitors, their keywords, keyword gap, position movers and overlaps, the result of the latest battle
BacklinksBacklink health, top referring domains, backlink gap, anchor distribution
BingSearch performance, top queries and pages, index coverage, backlinks

The agent keeps the brand terms of your pinned competitors out of its keyword recommendations, as described in chapter 6.

Sources come only from the library or a checked search

The agent does not write sources from memory, because that is how references to pages that do not exist come about. In the Content Studio, the content coach can search the web for sources instead and fetches every hit before it enters your text; such a search counts as five messages and is limited per day. Addresses that do not exist are removed, addresses that could not be checked are named so that you open them yourself. You can take found sources into the library as external sources.

Precise questions give better answers

1
"Which of my keywords have the best chance of reaching the top 3?"
2
"Where does my competitor example.com rank ahead of me?"
3
"What should I improve on my homepage to be cited by AI systems?"
4
"Which pages lost the most clicks in the last 7 days, and why?"

15Content Studio

Full step-by-step walkthrough: See Content Studio Step by Step for the complete tour — 6 entry points, 5 process steps (Library → Brief → Persona → Template → Editor), and the top 10 features no other content tool offers.

The library holds your own and external sources in three tabs

The library has two tabs. Under Brand knowledge / own sources you upload your own documents (PDF, DOCX, TXT, Markdown), import pages by URL or paste text; up to 50 documents per project. They are split into passages and made searchable so that the AI can draw on them. Under External sources you collect specialist articles, studies and documentation from other websites, up to 20 per project and free of charge; the content coach in the editor answers questions from the selected sources and names the source of every statement.

Active-for-AI multi-select: Each document has an Active checkbox in the leftmost column of the table. Only documents marked active feed into persona generation, brand keyword extraction, and the content-generation brand context. Inactive documents stay stored and can be re-activated any time. The header shows a counter „X of Y active for AI" plus „Activate all" / „Deactivate all" buttons that flip the entire set in one click. While a document is still being processed (status „pending" or „processing"), the checkbox is disabled — activation is only possible after the embedding completes.

The background for the AI names your most important page and two reference pages

In the library you set your most important own page and up to two reference pages, for example competitor pages, under “Background for the AI”. The editor uses them for keyword evaluation and comparison. The settings apply to all new drafts in the project.

A content brief prepares the article

Create AI-generated content briefs based on a topic, target keywords, and optionally a persona. The AI enriches the brief with your brand context from the active library documents. Each brief includes: title suggestions, H2 outline with key points, core messages, E-E-A-T recommendations, tone of voice, and target length.

Templates define the structure of an article

Build reusable structural templates for your articles in a drag-drop editor. 28 element types — TL;DR (Short Answer), Key Statement, Intro, Table of Contents, H2 Section, Definition Box, Step-by-Step, Comparison Table, Pro/Con, Recommendation by Profile, Statistic, Interactive Quiz, Pull-Quote, Key Takeaways, CTA, FAQ, Author Box, Author & Date, Source and the seven narrative blocks Chapter, Scene, Dialogue, Flashback, Turning Point, Resolution and Epilogues — combine via drag-drop into a template. 18 pre-installed default templates (Standard Article, Tool Comparison, Buyer's Guide, Listicle, How-To, Case Study, Tutorial with Quiz, Pillar Page, Study, Product Page, Landing Page, FAQ Page, Thought-Leadership Essay, Definition, News) are clone-able as starting points. In the brief, you pick the template from the gallery — the AI fills every element with topic-specific content in the right format. Templates are the reusable blueprint between brief and content: configured once, used for every sibling article in the same branch. See the dedicated „Templates" section further below for the full deep-dive.

The table of contents is built from the headings of the article

The table of contents exists as an element in the templates and as the Table of contents button in the editor. rankmio builds it from the H2 and H3 headings of the article: every entry jumps to its section, and H3 headings are indented below their H2. It only appears from two headings.

The table of contents is rebuilt on every save. If you change or delete a heading, the entry changes with it, and the table never lists a heading that no longer exists. The jump links are kept when the article is transferred to WordPress.

The author box shows the photo, role and profiles of the author

In the Authors tab you upload a photo for each author, as JPG, PNG or WebP up to 2 MB. rankmio crops it to a square.

The template element Author Box shows photo, name, role, short bio and profile links. All details come from the author settings, and the AI adds nothing. Without a photo, the box appears without an image, and if no author is assigned to the article, the box is left out.

The element Author & Date remains as a simple author line. In the editor, the Author button inserts either the author line or the author box and recommends the box when photo and bio are filled in. The box is rebuilt from the current author details on every save, so a new photo or a changed role appears with the next save. The photo also appears in the structured data, in the image field of the schema.org Person.

In the code the box is marked up as a complementary section: the photo as a figure, the name as “Written by” in an address element with rel="author" to the profile, and the profile links as a list named after the platform, such as LinkedIn, X or XING. In the field Areas of expertise you enter one topic per line. The areas of expertise and the short bio appear in the structured data of the article as knowsAbout and description and link the person to their topics.

Narrative mode writes prose instead of an article

Set Content goal to Story / narrative and the Content Studio writes prose instead of an article. The brief becomes a synopsis with characters, setting and dramatic arc; the article rules — TL;DR, key takeaways, FAQ, CTA, keyword density — are switched off, and the schema is marked up as ShortStory with the genre Belletristik so a machine does not read the text as a factual report. Two templates come with it: Short story and Narrative in chapters.

If an active library document names chapters, rankmio recognises them and offers to take them over. Everything the story needs — characters, plot, tone — comes from that document; it is handed to the model in full instead of being searched piece by piece.

An example with all settings: The first story rankmio told — briefing, synopsis, settings and the finished text.

Poem mode writes verse in a fixed form

Set Content goal to Poem and the brief becomes a poem plan instead of an outline. It sets the form, the number of stanzas and lines, rhyme scheme, metre, tone, a stylistic model if there is one, images and punchline. You can change each of these in the brief before the poem is written. Fixed forms such as sonnet, limerick, haiku or elfchen keep their measures.

The poem appears with a title, each stanza as its own paragraph and each line on its own line. rankmio counts stanzas and lines and has the poem rewritten once if the count does not match the plan. Template, research, placeholder images, sources and the citation readiness check are dropped, and the structured data marks the poem as a creative work with the genre Poetry.

In a poem your brand voice only sets the form of address, the words you have excluded and your word list, because form and model decide sentence structure and rhythm. rankmio only uses material about your company when the topic or instructions explicitly ask for it, for example a poem for your company's Christmas card.

Buyer personas describe your readers

Generate AI-powered buyer personas from your active brand documents or create your own. The AI identifies 2-4 personas with description, pain points, goals, preferred channels, and content preferences. You can also create personas manually or edit AI-generated ones.

Behaviour when no documents are active: If your library has documents but none is marked active, a confirm dialog appears before generation: „Create generic personas based on domain and industry topics?" — you consciously decide whether to proceed without brand context.

Persona names never appear in the article: The persona name (e.g. „Marketing Max") is an internal AI label only and is NEVER mentioned in the generated article text. The content addresses the audience generically by role („as a Marketing Manager", „many decision-makers") — so the article reads naturally instead of introducing an unknown character.

Every generated draft receives a citation readiness

Generate full SEO-optimized articles from your briefs. Each draft receives a citation readiness score (0–100) measured by the same rules as the GEO audit. Schema.org JSON-LD is shown as a copyable code block; the Mark up everything button in the schema card creates Article, FAQ and Quiz markup in one go, without AI and free of charge. Export your content as a complete HTML file with meta tags, Schema.org markup, and proper SEO structure.

Wikidata sameAs Enrichment

All Schema.org markup generated by the Content Studio (and the standalone Schema Generator) is automatically enriched with sameAs references to Wikidata and Wikipedia. This is one of the strongest signals available to AI search engines like ChatGPT, Gemini, Perplexity, or Google AI Overviews to disambiguate entities — e.g. distinguishing whether „Jaguar" refers to the car, the animal, or the sports team.

Narratives are the exception: with the content goal Story no entities are linked at all, because the settings and props of a story are not its topics. The matching happens context-based: rankmio compares your page content with Wikidata descriptions and selects the most fitting entity. Wikidata lookups are cached for 30 days for performance, and the schema is generated even if the Wikidata API is briefly unavailable (just without sameAs links in that case).

Why this matters: AI search engines rely on knowledge graphs to understand „who" or „what" your page is about. With Wikidata sameAs links, you give them an unambiguous anchor — significantly increasing the chance that your content is correctly attributed and cited in AI answers.

For a brief with research you choose the competitors first

When you create a brief with research or enter a search term in the editor, you choose the competitors whose pages rankmio analyses before the analysis starts. rankmio first shows the Google top 20 for your search term, in the language and country you set in the form. Both are preset with the project's language and country. You choose at least 5 and at most 10 pages. Forums, videos, shops and PDF files are not preselected because they rarely fit an article. You can add a page outside the top 20 by its address.

The list is free up to ten times an hour; after that, each further block of ten costs credits. rankmio then reads every chosen page, counts its terms and finds the search queries it ranks for. Each chosen page that has to be read costs credits; pages that are already available cost nothing. If a page returns no text for technical reasons, its price is refunded. A page in another language is charged even if it is not used for the terms afterwards. The button in the selection window shows the total price before anything is charged (see Credit prices). With “Continue without competitors” you create the brief without this analysis.

In the brief you manage your project's competitors

The brief form shows the five most important active competitors from your project's register. “Manage competitors” opens a window above the form. There you switch competitors on or off, remove them or add a new domain, and the form keeps what you have entered.

Every change applies immediately, including on the project page in the Content Studio and in the SEO area. Active competitors are marked and preselected first in the selection before an analysis. Inactive ones stay in the register but are not highlighted there. Accounts with only one of the two packages do not see the register in the Content Studio.

The brief form remembers what you entered

As long as a new brief has not been saved, your browser keeps what you entered. If you reload the page or come back later, the entries are back in the form, and a notice offers “Discard”. The saved state lasts seven days and only in this browser. Once the brief is saved, it is deleted.

You edit the text in a visual editor

The full WYSIWYG editor lets you edit generated drafts or write new content from scratch. Features include auto-save (every 60 seconds) and URL import for existing pages. If you open the editor with a page address, for example from a task, the page comparison, the click-through rate or Search Console, it loads the text of that page; a notice shows while it is loading. Opening the same task a second time continues in the existing draft. The editor opens at 900px height (+50% vs. before) and can be resized further via the drag handle at the bottom.

No phantom drafts: Opening the editor briefly and leaving without entering content does NOT create an empty „Untitled" entry in the drafts list. The draft is created only when you actually do something — type a title (more than 2 characters), write text (more than 5 characters), insert a media element, or explicitly click „Save".

The editor checks citation readiness while you write

The right-hand panel checks the draft by the rules of the GEO audit while you write, free of charge: one H1, at least two H2 headings in longer texts, a clean heading order, short paragraphs, lists or tables, and for posts author and date. Open points are at the top, everything already fulfilled can be collapsed. Below that come content hints that do not count towards the score: schema, factual claims, clear statements, sources and the search term in headings and text. Only the assessment of named entities uses AI and costs 1 credit.

Many content management systems output the title of a post as H1 themselves. In that case you tick My CMS sets the title as H1 once per project. rankmio then checks that the text itself contains no H1; without the tick, the text needs exactly one.

The panel follows the order of work. If you opened the editor from a task, the Action required box is at the very top, followed by citation readiness, internal links and schema.

The Internal links box shows which links the page should set according to the GEO audit, which pages should link to it, and which pages of your website fit the topic of the draft. Insert places the link at a suitable spot in the text. The box opens by itself when the audit has recommendations for this page and is collapsed otherwise.

The content coach reads the whole draft

The chat in the editor works as a content coach. It reads the whole draft, including long articles, and so also finds contradictions and repetitions between sections far apart. Its suggestions appear as cards that you apply with one click. If you opened the editor from a task, the Action required box checks on request whether the task is already solved in the draft. For tasks that cannot be measured in the draft, such as a better position, it says instead that the effect shows in Search Console after publishing.

Images and videos come straight into the text

Every generated draft automatically receives default image and video placeholders inserted at H2 positions (configurable per brief, default: 1 image + 1 video). You can swap them with one click via the Media tab in the editor sidebar, which offers three sub-tabs:

  • AI image generation — Generate custom images via OpenAI gpt-image-1.5 or Google Gemini 3.1 Flash Image. Free-text prompt, three sizes (1024×1024, 1024×1536, 1536×1024), and optional alt text. Files are stored on rankmio and served via an authenticated endpoint.
  • Pexels stock photos — Search the Pexels library (free, commercially usable, no attribution required). Toggle «Show photos with people»: by default people-tagged images are hidden, activating the toggle requires explicit consent. Inserted images use the direct Pexels CDN URL, so they work in exported HTML on any host.
  • YouTube videos — Search the YouTube Data API v3. By default only Creative-Commons-licensed videos are shown; with the „All licences" toggle you can also embed regular videos (with a responsibility note for your use case). Inserted videos use the privacy-friendly youtube-nocookie.com embed.

Swap an existing image or video in place: Click the existing image or video in the editor (it gets an orange outline as a visual marker), then switch to the Media tab and click a new search card or generate a new AI image → the marked element is replaced instead of appended at the end. Works with all three sources (AI image, Pexels photo, YouTube embed) and with the auto-inserted default placeholders.

When you save the draft, all inserted media is stored as part of the HTML and exported with the article. Cursor-position inserts and automatic nested-figure flattening keep the markup clean.

rankmio marks the places where an existing page can improve

You revise an existing page with the “Revise page” button. It appears next to a recommendation in the Citation Tracker and under AI recommendations in the Content Studio whenever the recommendation concerns one of your existing pages. rankmio loads the page into the editor as a draft and checks it, which usually takes half a minute to three minutes. The hints for the revision then appear next to the text.

rankmio does not change the text itself. It marks places and says why a change would help there, for example because evidence or a source is missing, a section does not answer the question, or something already appears further up. “Show passage” jumps to the place in the text, and “AI suggestion” writes a suggestion for that one place only, which you adopt only after checking it. Checking and suggestions cost no credits. Your website only changes once you publish the draft or write it back to WordPress.

The editor underlines misspelled words

While you write, the browser checks the spelling and underlines misspelled words in red. Right-clicking the word shows suggestions. The dictionary follows the language of the project, not the language of the interface. The check comes from the browser and hardly covers grammar or style. You edit tables and links via the toolbar and the Table menu.

The term bar shows what the top pages write

Once you have chosen competitors, rankmio reads their pages and shows the term bar in the editor. A ring shows the score of your text, half from the terms and half from citation readiness, and next to it the target set by the best chosen competitor.

The terms appear as chips with their state: missing, too few, within range or too often. The range is scaled to the length of your text. Clicking a chip highlights where the term appears in your text and inserts nothing. Below are the search queries the top pages rank for and, on request, their subheadings, which you can insert one at a time at the cursor as a prompt for your own wording. You change the basis via “from n pages · change”, and recalculating costs nothing.

If one of your own pages already ranks at position 11 to 30 for these terms, the bar names it with “Load this page”, because reworking it usually pays off more than writing a new one. Without a brief, the empty editor asks “What are you writing about?” and starts the same selection with your search term. The analysis runs before you save, and a draft is created only when you save.

The keyword panel collects terms from three sources

The sidebar Keyword Panel extracts keywords from three sources: your own page, competitors, and brand documents. Keywords can be selectively enriched with real SERP data — search volume, keyword difficulty, position, and search intent. The panel features sortable column headers (Keyword, Volume, Position, Intent, Occurrences), live occurrence counting in the editor text, keyword stuffing warnings, AI keyword extraction, and the ability to delete or insert keywords with one click. Costs: see Credit prices.

The editor sidebar has six tabs

Below the checks, the editor sidebar has 6 tabs in this order: Keywords, Brief, Sources, Topic gaps, Compare and Media. Notification badges show pending action items:

  • Keywords — 4 source tabs (Own/Competitor/Brand/Project) × 3 sub-tabs (Rankings/Content/Keyword gaps). Live occurrence counter, GSC enrichment, click-insert. The „Content" sub-tab always extracts keywords from the current editor text (prevents stale cache when switching between drafts).
  • Sources — Top SERP URLs + People-Also-Ask questions (click-insert as H3) + SERP features. At the top sits the collapsed box “Load Page Content”; it loads a page by its address or, with a WordPress connection, a post straight from WordPress.
  • Topic gaps — Research-based content gaps with status traffic-light (red/yellow/green), Insert/AI-Generate/Save-Keyword per gap, progress bar.
  • Compare — citation readiness bar comparison + Word count comparison (your article vs. competitor average as bars) + Per-criterion comparison: list of citation readiness points with your own Pass/Fail (green/red dot) + competitor quota (X/Y pass) — red number when majority pass and you don't.
  • Media — AI image (OpenAI gpt-image-1.5 / Google Gemini 3.1 Flash Image), Pexels stock search, YouTube search (CC-license filter), click-to-replace existing media.
  • Brief — Read-only view of the original brief: topic, keywords, persona, template, goal/tone/length, title suggestions, H2 outline with key points, core messages, E-E-A-T recommendations, guardrails. Visible only when the draft has a brief.

Notification badges (small colored numbers in tab corners): Keywords (amber) = keyword-stuffing warnings (density > 3 %); Topic gaps (red) = uncovered gaps; Compare (amber) = competitors with higher citation readiness; Media (pink) = default placeholders not yet replaced.

A progress bar shows the generation

When you click „Generate content", a progress bar shows the working steps, from preparing the brief to the citation readiness check. It reaches 95 % in about 60 seconds and only jumps to 100 % when the text is really finished. Then the editor opens.

Several checks keep invented content out

Five checks on the server make sure that a text contains no invented sources, links or figures.

  • A text may only name sources that really exist: addresses from the research for the draft, from the active library documents and from the brand context. If the AI still invents a link to a subpage, such as an article on ahrefs.com, rankmio turns it back into plain text. The homepage of a well-known organisation stays allowed as a link.
  • If an active library document contains a product list with shop addresses, rankmio links every product mentioned in the text to its shop page, also in tables, FAQ answers and lists. It recognises shortened product names too and corrects links that point to the wrong shop.
  • The AI fills a statistics box only when the figure and its source come from the brand context, the research or the library. Otherwise it leaves the element out.
  • If the library or the persona changed after the brief was worked out, rankmio works the brief out again with the new state before writing. No text is written with outdated knowledge about your brand.
  • For citation readiness the AI may use generally known figures, but a specific statistic only with its source. The text also names at least five concrete things, such as tools, platforms or methods, instead of generic terms.

One template per article type carries a whole series

Templates define the structure of your articles — which elements appear in which order. They are the reusable blueprint between brief and content. Instead of writing each article from scratch with a different shape, you configure ONE template per article type and reuse it for every sibling article in the same branch.

The 28 element types are listed in the section on templates above.

18 pre-installed default templates, tagged for content goals (traffic / leads / conversion / authority / education / support):

  • Standard Article, How-To / Guide, Tool Comparison, Buyer's Guide, Listicle ("X best …")
  • Case Study, Definition / Glossary Entry, News / Update, Pillar Page, Tutorial with Quiz
  • Study / Data Report, Product Page, Landing Page, FAQ Page, Thought-Leadership Essay

Workflow: open Content Studio → Templates tab. Standard templates are read-only — clone one to get an editable copy in "My Templates". In the editor: drag-drop reorder, toggle elements on/off, set hints per element (topic-specific instructions for the AI), configure FAQ min/max and table dimensions, set default target length + tonality + recommended goals. In the brief form, the template gallery shows recommended templates first based on your selected content goal.

Why this matters for scaling: for Semi-Programmatic SEO (a series of "Tool A vs Tool B", "X best Y" or similar articles), you configure ONE template once. Every sibling article in the branch uses the same template — structurally identical, content-wise fresh. The AI receives the template as a mandatory structural blueprint and fills topic-specific content into each element.

Schema.org per element: when FAQ is in the template, FAQPage schema is auto-generated; HowTo for step-by-step; Quotation for pull-quotes; Table for comparison tables; DefinedTerm for definition boxes. Plus the Wikidata sameAs enrichment for the main entity.

Cost: Templates feature is free. Cloning + editing are free. The brief generation that uses the template costs the normal Content Brief credits.

Briefs can take over competitor keywords

When you create a brief, the keyword dialog offers three sources: AI suggestions, your project keywords and the data package of your pinned competitors. Only pinned competitors appear there. Your selection is kept across all tabs of the dialog, and the AI suggestions only start when you click.

Every project starts with two standard briefs

Each new project receives two read-only standard briefs, an argumentative essay and a standard specialist article. You clone them and edit the copy freely; the standards remain unchanged as a reference.

You can dictate text into the editor

Click into the text so the cursor is active, then on the red microphone in the toolbar; the browser asks for permission to use the microphone. Speak for up to 60 seconds and click the microphone again. The recording is transcribed, inserted at the cursor position and then discarded. One dictation per day is free; the label next to the microphone shows beforehand whether the next one costs credits.

The schema card checks the markup of the draft

The schema box in the editor sidebar shows in its header which markup the draft carries. It is collapsed and opens by itself when the source page contains several conflicting blocks or when the FAQ in the text and in the markup differ. If the text contains at least three question-and-answer pairs but no FAQ markup, it points this out. The Mark up everything button then creates Article, FAQ and Quiz markup without AI and free of charge.

Your schema builds on the markup of the source page instead of replacing it. Publisher, logo, topics and links between the blocks are kept; headline, description and date come from the draft. The comparison window shows the original, your schema and a table that sets both side by side, field by field.

Drafts go straight to WordPress

From every draft in the editor you can push directly to your WordPress site — no copy-paste, no manual HTML upload. Set up the connection once under Profile → Publish: per site you save a label, the WordPress URL (HTTPS mandatory), the WordPress username and an Application Password. The Application Password is generated in your WordPress profile under „Application Passwords" (WP 5.6+) and stored AES-256 encrypted in our database. You can revoke it any time in WordPress without touching your main login.

Multi-site: One account can manage multiple WordPress connections — useful for agencies handling several customer sites. Each connection is independent (own credentials, own test-status).

Publish modal: Click „Publish article" in the editor top bar. After picking the connection the modal asynchronously loads the WP site's categories, tags, post types and detects whether Yoast SEO is active. You then choose:

  • Post type — Post or Page (custom post types are listed too if registered for REST).
  • Categories — Pill-style multi-select with QuickAdd: type a name + Enter → category is created in WordPress and selected.
  • Tags — Same pill-style multi-select with QuickAdd.
  • Featured image — First image in the article is auto-extracted and uploaded to WordPress media library; an optional URL override lets you supply another image.
  • Slug + Excerpt — Both optional, auto-generation falls back to the title / first 240 chars of content.
  • Yoast SEO — If Yoast is detected on the WP site, an extra section appears with SEO title (60 char counter) and meta description (155 char counter) — both stored in the Yoast block.
  • WP status — Save as draft (default, recommended) or publish immediately.
  • When — Now or scheduled for a specific date and time.

Schema: If the WordPress site already marks up its posts itself, for example with Yoast, Rank Math, All in One SEO or SEOPress, rankmio only transfers the additions such as FAQ or quiz and does not add a second article markup. After publishing, rankmio checks whether the schema has arrived on the page and reports the result. Some sites remove scripts from the post content; the message then says so.

Scheduled publications wait in a queue that rankmio works through every two minutes. If the connection to WordPress fails for a moment, rankmio tries again up to four times, after 1, 5, 30 and 120 minutes. Under Profile → Publish you see scheduled and past publications with their status and planned time, and after success the link to the post in WordPress.

The connection to WordPress has to use HTTPS, because the application password would otherwise travel unencrypted. rankmio rejects addresses from local or private networks such as localhost, 10.* or 192.168.*. Every publication is logged per user.

Publishing to WordPress costs no credits.

You load existing posts from WordPress and write them back

In the editor, open the box “Load Page Content” in the Sources tab and choose “From WordPress”. The list shows the posts and pages of the connected site, most recently modified first, with status and modification date; the search field narrows it down. “Load” creates a new draft in the current project. Nothing changes on the site.

The editor shows paragraphs, headings, lists and quotes as text. Images, galleries, columns, embeds and plugin blocks appear as a line of their own such as “⟦WP-Block 3 · image⟧”. Leave these lines in place: when writing back, rankmio puts the unchanged block in their position. If you delete a line, the block will be missing from the post afterwards.

Under “Publish article”, such drafts let you choose between “Update the existing post” and “Create a new post”. When updating, title and text replace the content of the post; address, status, categories, featured image and SEO settings stay as they are. Beforehand, rankmio compares the draft with the current state in WordPress and shows every changed paragraph in its old and new version, along with title changes and missing blocks. Paragraphs you have not changed go back character for character as they came. A post from the classic WordPress editor stays classic; if anything in it changes, its text comes back newly formatted and looks the same.

If the post was changed in WordPress after you loaded it, rankmio warns you. The “Update post” button then stays locked until you explicitly confirm the overwrite, because the changes made in WordPress would be lost. After updating, “Undo” reverses the step as long as nobody has changed the post in WordPress since. Independently of this, WordPress saves a revision before every overwrite.

rankmio recognises posts built with page builders such as Elementor, Divi, WPBakery, Beaver Builder or Oxygen. You can load and check them but not write them back: their content is not stored in the WordPress text field, and writing back would break the layout. Loading and updating cost no credits.

Title and meta description are written back once your website releases them

rankmio reads the title and meta description of your posts from Yoast SEO, Rank Math or SEOPress and shows them in the editor. It can only write them back once your WordPress website releases these fields for its interface. A short piece of code is enough for that; you do not need a rankmio plugin.

  1. Install the free plugin “Code Snippets” in WordPress if you do not use it yet. Alternatively you can paste the code into the functions.php of your child theme.
  2. Under Snippets → Add New, create a snippet “rankmio SEO fields” and paste the code. The plugin adds the line <?php itself; if you paste it a second time, the snippet cannot be activated. Below, choose “Run snippet everywhere”.
  3. Click “Save Changes and Activate”. In the list under “Snippets” the switch is then set to active.
  4. Load a post from WordPress in the Content Studio. The box “Title and meta description” then says that rankmio writes both back when you update the post.
/**
 * rankmio: Title und Meta-Description über die WordPress-Schnittstelle lesen und schreiben.
 * Gilt für Yoast SEO, Rank Math und SEOPress, für Beiträge und Seiten.
 */
add_action('init', function () {
    $felder = [
        '_yoast_wpseo_title', '_yoast_wpseo_metadesc',      // Yoast SEO
        'rank_math_title', 'rank_math_description',          // Rank Math
        '_seopress_titles_title', '_seopress_titles_desc',   // SEOPress
    ];
    foreach (['post', 'page'] as $typ) {
        foreach ($felder as $feld) {
            register_post_meta($typ, $feld, [
                'type'              => 'string',
                'single'            => true,
                'show_in_rest'      => true,
                'sanitize_callback' => 'sanitize_text_field',
                'auth_callback'     => function () { return current_user_can('edit_posts'); },
            ]);
        }
    }
});

The code only allows reading and writing for accounts that may edit posts in WordPress, which includes the Editor account you created for rankmio. It does not work for All in One SEO, because that plugin stores its details differently.

These credits and limits apply

First document upload and first URL import per project are free. Pexels stock photo search, YouTube video embeds and WordPress publishing are free. The rule-based citation readiness check is always free. All other Content Studio actions cost credits — exact prices: see Credit prices.

16Monitoring

This chapter describes the tools that measure regularly and show how your backlinks, your rankings and your visibility in AI answers develop. Most of them sit in the Monitor group of the sidebar.

The Backlink Hub loads the backlinks of your domain from DataForSEO and spreads them over six tabs: overview, top backlinks, anchor texts, referring domains, history and competitor gap. The main figure of the overview is the backlink strength from 0 to 100: it counts every referring domain by its quality, spam does not count, and domains confirmed by Google or Bing weigh more. The visibility index uses this value. Below it the health score gives a second assessment from domain variety, spam share, anchor texts, trend and follow ratio; it does not count in the index. Each link shows its spam score as a measured value from high to unremarkable; rankmio does not judge any link as harmful.

The number of referring domains is the same everywhere: rankmio counts the domains from the link database and from your uploaded Google and Bing link reports, each only once and without spam. The tiles say how many domains and backlinks come from your own projects and how many spam domains were not counted. The Rank column shows the authority of a domain on the DataForSEO scale from 0 to 1,000, not the Ahrefs Domain Rating from 0 to 100. The History tab shows one value per month.

Under Findings in the link profile rankmio describes recurring patterns, such as the same anchor text on several domains or a cluster of new domains within a few days. A pattern is an indication, not proof. From these groups the link profile check builds a disavow file on request.

Besides the data from DataForSEO you can upload exports from Search Console and the Bing Webmaster Tools in the field Upload data from Google and Bing at the top of the hub, several files at once. rankmio recognises the type of report and the separator itself. From the Search Console you upload the links report as a ZIP archive or as a single CSV; from Bing the backlink reports, the most detailed being Referring Pages. A table then shows for each row whether it was found by both sources or by only one. Neither export replaces the data from DataForSEO, because neither names the follow status or a spam score.

The competitor gap shows domains that link to your pinned competitors but not to you. You choose between a quick pass with 200 domains per competitor and a thorough one with 1,000. The filters show domains from rank 100 or from rank 200; the Hot figure counts domains from rank 100 that link to at least two of your competitors. Suspicious domains and endings from other markets are hidden from the start; you can show them again.

The position tracker follows your rankings

The position tracker fetches the ranking keywords of your domain in one bulk request instead of checking each keyword individually, up to 200 in the starter configuration and up to 1,000 in the larger ones. You choose on which weekdays it measures. The share-of-SERP score combines the top-3 share (40 per cent) and the top-10 share (60 per cent) of the last seven days. A quick-wins block shows keywords on page two, positions 11 to 20, with high search volume. A watchlist adds individual terms, and the larger configuration compares you with a competitor.

The citation tracker measures your presence in AI answers

The citation tracker asks AI systems such as Google AI Overviews, Gemini, Perplexity and ChatGPT your questions and records whether your domain is cited or merely mentioned. ChatGPT is queried through its real interface. All questions of a project are kept in one watchlist; each shows where it came from and whether it is about your brand or a topic, and you can filter, edit and remove several at once. You choose how often it measures: every week, every second or every fourth week. New questions are suggested by the phrase discovery, or you enter them yourself; scanning and taking over can be chosen separately, and taken-over questions bring their results into the tracker.

The overview shows the citation rate, the comparison between the AI systems and the strongest phrases; the history shows how the rate develops. For phrases where you are never cited, rankmio suggests creating a brief or improving an existing page. Queries that could not be carried out are shown as "not measured" and are not counted.

The phrase discovery finds the questions for the citation tracker itself

The questions you give the citation tracker determine what it measures. Instead of making them up yourself, you let the Discovery tab suggest them. rankmio derives them from data that already exists about your website and then checks them in the AI systems.

First you choose what you want to know. Brand visibility is about whether an AI mentions you by name, for example when someone asks about your brand, about alternatives or for a comparison. Topic visibility is about whether it cites you as a source although your name does not appear in the question. The second tells you more about whether new readers find you through an AI.

SourceWhere the questions come fromApplies to
Bing Copilot queriesQuestions users asked Copilot where your site was cited. Requires the uploaded Bing AI Performance report.Topic
Pages from Google's AI answersGoogle search queries for the pages Google itself shows in AI answers. Requires a connected Search Console.Topic
Search Console queriesReal Google searches your website is found for, without your brand nameTopic
AI topics from your pagesTopic questions without your brand name, derived from the titles and headings of your pages. This also works without Search Console.Topic
FAQ from structured dataQuestions you have added to your pages yourself as FAQ markup (FAQPage)Brand and topic
Content topicsQuestions your most important pages answerBrand
BrandQuestions about your brand, about alternatives and for comparisonsBrand
Competitor listsQuestions such as "the best tools for …" whose answers name competitorsBrand

In guided mode you choose the sources yourself. In automatic mode you only say how many questions you want, between 5 and 50, and rankmio mixes the sources. The page shows what a suggestion costs before you start; the first run with Search Console queries is free.

You can edit or delete the suggested questions or add your own. Then you choose which questions are checked and in which AI systems: Google AI Overviews, Gemini, Perplexity, ChatGPT and Claude. The live scan shows for each question whether your domain was cited. Queries a provider did not answer count as not measured and are not included in any rate.

What you want to watch permanently you transfer to the watchlist. The results of the first scan come along and are not charged again. A question that is already in the watchlist in the same or almost the same wording is not added twice. From then on the citation tracker keeps measuring at the interval you have set.

Bing AI Performance shows real Copilot citations

The Bing Webmaster Tools contain a report on how often Microsoft Copilot cites your pages and with which internal search queries it found them. Microsoft offers no interface for it, so you upload the export yourself.

1
Open bing.com/webmasters and select your website.
2
Click AI Performance in the left sidebar and choose a period of 30 or 90 days.
3
Export the report as CSV and upload it in rankmio under GEO → Measured citations → Bing Copilot.

rankmio accepts the overview with its daily figures and the detail table of cited queries, in English or German, separated by comma, semicolon or tab, up to 10 MB and 50,000 rows. The dashboard then shows the trend over 90 days, your most cited pages, the most frequent queries, and their intents and topics. The figures feed the actual citations in the GEO score.

In the same way, GSC AI Performance shows impressions in Google's AI answers from an export of the Search Console.

Traffic recovery shows where you are losing clicks

Traffic recovery compares the clicks of each page in the last 28 days with the 28 days before and names the cause the data suggests: weaker rankings, lower demand, a snippet problem or a lower click rate without a clear cause. Below each page is the query that lost most. A loss counts from two clicks; pages with fewer than ten clicks before are collapsed below, because chance is possible there. The note on new queries only appears where they explain a worse position.

Below that are the queries on page one where a better search result text brings more clicks, with a link to the snippet tool. Queries in positions 11 to 20 are in the Keyword Hub under Opportunities.

A weekly email summarises the project

Every Monday the account owner receives an email per project with the tasks completed in the past week, the most important open tasks, the development of the Copilot citations and the AI champions.

For consumer purchases, a 14-day right of withdrawal applies. The exact procedure is explained in our Right of Withdrawal information — including the online withdrawal form.
FAQ Privacy Policy Terms Withdrawal Imprint

Frequently asked questions about the manual

The rankmio Score is a weighted overall score from 0 to 100 built from eleven analysis pillars — among them PageSpeed, the SEO checklist, keywords, the GEO score, Core Web Vitals, sitemap and robots.txt. It appears in the dashboard below the Analysis Status. The prominent figure on the dashboard is the visibility index, which is calculated differently.
The GEO audit crawls up to 400 pages of your website and calculates the GEO score: half citation readiness, that is the prerequisites measured on the HTML of each page so that an AI can cite it, and half actual citations, that is how often AI systems demonstrably cite your pages. The results name the open requirements per page and where the bottleneck lies.
The rankmio Chat Agent is an AI-powered SEO and GEO advisor with direct access to your project data. It reads the measurements of your project with read-only tools and suggests the matching tool where a new analysis would help; it does not start analyses itself.
Credit costs vary by feature. Many features like the dashboard and reports are free, keyword enrichment is cache-aware (cached keywords cost nothing). The complete, always up-to-date price list per function is published on the credits page at /credits#feature-prices.
Citation readiness shows whether a page meets the requirements for AI systems to use it as a source. It checks a page by the same rules as the GEO audit, such as one H1, a clean heading order, short paragraphs, lists, author and date. Content hints such as factual claims, clear statements and sources are shown alongside but do not count towards the score.