For a small website, SE Ranking mainly delivers the data that Google Search Console fundamentally does not have: search volume, keyword difficulty, competitors' terms and positions, and other sites' backlinks. That is what my test of 17 October 2026 shows, in which my AI employee for SEO did the same seven tasks once without and once with SE Ranking, connected through the Data API and the MCP server. For your own positions, though, the tool brings nothing better: the database knew 19 of the 8,144 search terms for which my pages appear in Search Console, and the rank tracker did not find 37 of 50 terms in the top 100, including 28 for which Search Console shows positions. The whole test run cost 19,155 credits. Without a platform subscription, the API is available pay-as-you-go from 45 euros per 250,000 credits with no annual commitment; SE Ranking confirmed this to me in writing (as of 20 October 2026). My verdict: a clear win for volume, competitors and backlinks, while for your own site Search Console remains the source.
Upfront: I have a partnership with SE Ranking, and I earn a commission through my links to SE Ranking*. SE Ranking had offered me a partnership; I still ran the test first, with the free trial I had set up myself. Only afterwards did we agree on the partnership: for this article and three others, SE Ranking provides me with an account free of charge. SE Ranking had no influence on the setup, results or verdict of this test and did not see the text before publication.
* Affiliate link: if you subscribe or buy credits through this link, I earn a commission. Your price stays the same. I have a partnership with SE Ranking.
The tester was Sebastian, my AI employee for SEO and GEO, with a trial account of 14 days and 100,000 credits. Which free tools formed his toolbox beforehand and where a paid suite comes in at all is covered in SEO Audit Tools Compared.
What SE Ranking is and what I tested
SE Ranking is an SEO platform with a rank tracker, website audit, keyword and competitor database, backlink index and a section for visibility in AI search. I did not test the web interface but the access an AI employee needs: the Data API for large queries and the MCP server, through which Claude calls the tools directly. The MCP server reported 236 tools in four groups for us, 66 for data, 127 for projects, 28 for the AI section SE Visible and 15 for local SEO. SE Visible was not included in the trial account, and queries to it ended with an authentication error.
How to set up the MCP server in Claude Code without storing the key in plain text is described in my article Claude SEO: SEO With Claude Code, Skills, MCP. Here the question is what comes out of it.
How the test was set up
I wanted to know how much better an AI employee works with SE Ranking than without. So Sebastian did the same seven tasks twice, on the same evening:
The first pass ran without SE Ranking and used only free sources: Search Console, Bing Webmaster Tools, my own crawl, Lighthouse, web search. The SE Ranking MCP server was technically removed from the session in this run, not just forbidden by instruction; the REST interface was ruled out by the working instruction. Only after this pass was finished and secured with checksums over all 71 files of the first pass (results, raw data, scripts) did the second pass with SE Ranking start. At the beginning and at the end of the second pass, all 71 checksums matched, so the first pass was not changed afterwards. How the case study ran in detail is described in AI Employee With and Without an SEO Tool.
A finding only counted if it was confirmed against a primary source, meaning Search Console or the live page.
The results per task
Positions: Search Console clearly wins
SE Ranking counts 19 search terms for kevinwelter.com in the German database, while Search Console counts 8,144 in the 28 days to 16 October 2026. That is 0.23 percent. SE Ranking estimated the traffic at 5 visits a month, while Search Console showed 1,206 clicks for Germany in 28 days. For a small, young website this is to be expected: a third-party database mainly knows terms with measurable volume, while Search Console knows every query for which the page was shown.
The rank tracker, by contrast, measures specifically the terms you give it. Of 50 terms, it had a position for 13. For the 11 terms for which both sources had a value, it was off from Search Console by 7.6 places on average. It did not find 37 terms in the top 100, including 28 for which Search Console shows positions between 2 and 58, including “seo für kleine unternehmen” (SEO for small businesses) with an average position of 11.1. I have not worked out the cause. Part of the deviation is method: Search Console averages across devices, locations and days, while the tracker measures one moment in one place. That does not reliably explain the 28 total misses, though. Until a daily series clears them up, I use only Search Console for my own positions.
Search volume: the biggest gain, for little money
Without SE Ranking, Sebastian had no volume for a single search term. With SE Ranking, he sent the 1,000 terms with the most impressions in one query. 411 of them got a volume greater than zero, and the remaining 59 percent were mostly long, question-like queries without measurable demand. 300 credits were charged. According to the credit table in the docs it would have been 10 credits per term, so 10,000.
The numbers are rough for small terms. 89 terms had more impressions in 28 days than their monthly volume allows, for example “claude kosten” (Claude costs) with 4,953 impressions at 2,900 searches a month. The relationship between volume and impressions is loose (rank correlation 0.34), which is to be expected: volume measures demand, impressions measure your own visibility. Still, this is the part that has changed my work the most. Since then, I also read the monthly trend of a term in every prioritisation, because individual values jump from 10 to 20 up to 590 within a few months.
Gaps compared with competitors
Without a tool, Sebastian could describe only three qualitative gaps. With SE Ranking, the rule set defined in advance (competitor at position 20 or better, my website not) produced 51 gaps that fit the business. He checked 15 of them against Search Console and the live page: 13 were valid, 6 clearly actionable. These included “seo vs geo” with 480 searches and difficulty 11, for which my GEO article had only three sentences, and “ki agenten bauen” (building AI agents), for which five of my own pages compete with each other.
The number is a lower bound. In the trial account, every domain query returned at most 10 rows regardless of the limit set, and not more even via a second page. That is not in the docs. Sebastian therefore split the queries into topics and volume ranges. According to SE Ranking, the limit does not apply to paid accounts; I have not measured that yet.
Backlinks
Bing did not know a single incoming link, and Search Console does not deliver links through the interface. SE Ranking found 41 referring domains with 56 links. Around 30 of them are short-URL or spam sites without editorial context. Sebastian confirmed live three editorial links from other websites, all to English articles, plus book reviews on a professional portal. For the four competitors, backlink metrics with conspicuous anchor texts and spam shares were added. Without a paid tool, this view does not exist.
Technical audit
I started the audit with the lowest crawl rate, one request per second, because my website rejects too many requests in a short time. 554 pages took 12 minutes and cost 1,108 credits, 2 per page. Result: one shared finding with my own crawl (two page pairs with the same title in German and English), two small new findings on the privacy page, and three false alarms, such as “links without anchor text” for icon links with aria-label. New and useful were the field data from the Chrome UX Report: on mobile, 93 percent of page views load the largest content in good time.
The audit that ran at standard speed when the account was set up, by contrast, reported 18 errors instead of 4, including pages with 4xx errors that all answered with 200 live. If you point SE Ranking at a website with rate limiting, turn the crawl rate down and check errors from a fast run live.
AI search in Germany
SE Ranking listed kevinwelter.com with 28 prompts in which the AI answer links to my pages, all in Google's AI Overviews, none in ChatGPT, Perplexity, Gemini or AI Mode. 15 of them were English variants of “website security check”, and none concerned Claude, AI employees or SEO. For comparison, I queried a large German competitor domain: 1,455 prompts in AI Overviews, 66 in Perplexity, 43 in ChatGPT, 39 in AI Mode, 25 in Gemini. For Germany, the database is essentially an AI Overviews database. For ChatGPT and Perplexity, a small site still needs its own prompt set.
This block was the most expensive: 9,800 credits, of which 7,500 alone went to a ranking across five domains. For one competitor, the brand recognition counted a generic word as a brand and arrived at 4,622 mentions; that number is not reliable. The individual prompt queries, 200 credits per prompt according to the list, were not charged in the test. According to SE Ranking, that is not the norm; plan for the 200 credits.
Briefs
For three existing articles, Sebastian wrote briefs in both passes. With SE Ranking, volume and difficulty per target term were added, plus six new terms with demand, for example “website bei google eintragen” (add a website to Google) with 210 searches. The database delivered hardly any question data for German. From both briefs, an optimisation run on three pages came out on 17 October 2026; whether it works, Search Console will show after 28 days.
SE Ranking pricing: what the test would have cost
Pay-as-you-go and the annual packages are two separate routes. When I asked, SE Ranking confirmed in writing on 20 October 2026 that pay-as-you-go from 45 euros per 250,000 credits comes with no annual commitment. At that price, my whole test run of 19,155 credits would have used up about 3.50 euros of credit; the minimum deposit is 50 US dollars. If you add the prompts that were not charged at list price, it would have been 32,955 credits, just under 6 euros. According to SE Ranking, prices and limits change regularly; the current ones are on the pricing page of SE Ranking*.
On the credit logic itself: you are billed per query and per row delivered, and only successful responses cost anything. In the test, the table matched exactly for backlinks and audit, but not for volume and prompts. I therefore measure costs against the balance before and after each block, not against the list. In the test, rank tracker and position checks from the project tools used no credits but subscription limits.
Where SE Ranking hits its limits
- Coverage of your own site. 19 of 8,144 terms. For your own visibility, the free Search Console is the better source.
- Rank tracker unresolved. 37 of 50 terms without a hit in the top 100, of which 28 despite a Search Console position. In my test, not in general; the cause is not resolved.
- 10 rows per domain query in the trial account. Undocumented, slows down every competitor analysis. According to SE Ranking, the limit does not apply to paid accounts.
- False alarms at standard speed. On websites with rate limiting, throttle the crawl first.
- AI search Germany almost only AI Overviews. No substitute for your own measurements in ChatGPT, Perplexity and Copilot.
- Billing differs from the table. In the test, volume and prompts were billed differently from the credit table. According to SE Ranking, prices and limits change regularly.
Who SE Ranking is worth it for
For my needs, it is occasional, not monthly: before a topic plan, before a competitor analysis, before a revision. So I would take pay-as-you-go rather than a subscription. How this fits into a workflow is shown in How to Run an SEO Audit With AI.
Frequently asked questions
How much does SE Ranking cost?
Without a platform subscription, the API is available pay-as-you-go from 45 euros per 250,000 credits, with no annual commitment and a minimum deposit of 50 US dollars. The platform costs 109 euros (Core) or 235 euros (Growth) per month, 20 percent less when paid annually; for more API volume there is an add-on from 39 euros per month with annual billing. The Data API docs also mention annual packages starting at 12 million credits. As of 20 October 2026; according to the vendor, prices change regularly.
Is SE Ranking worth it for a small website?
For the question of how many people search for something and who ranks at the top for it, yes. Volume for 1,000 terms cost 300 credits in the test. For your own positions, no: the database knew 19 of 8,144 search terms of my website, and the free Search Console is enough for that.
How accurate is the SE Ranking rank tracker?
In my test, for 11 comparable terms it was off from Search Console by 7.6 places on average. It did not find 37 of 50 terms in the top 100, including 28 for which Search Console shows positions. The cause is not resolved. Until a daily series clears this up, I use the tracker only as an observation.
Can I use SE Ranking with Claude?
Yes. SE Ranking runs an MCP server that you add in Claude Code with claude mcp add, either by logging in through the browser or with an API key in the header. According to the vendor, a key has the same rights as the account, and read-only keys do not exist yet. That is why I block deleting tools in Claude Code.
Does SE Ranking replace measuring AI visibility?
Not for Germany. The database listed my website only in Google's AI Overviews, not in ChatGPT, Perplexity or Gemini. Even a large German comparison domain had only 25 to 66 prompts per AI service there. For these services, a custom prompt set remains necessary.
Where to go from here
Before you buy a data tool, set up the free sources and write down which question they cannot answer. For me it was volume, competitors and backlinks; that is exactly where SE Ranking delivered. The free foundation is described in SEO Audit Tools Compared, and how an AI employee works with MCP servers is explained in Set Up a Claude MCP Server. How Sebastian is built and how you can hire such an employee yourself, I show in my community Claude Practitioners.
If you want to try SE Ranking with your own question: the trial at SE Ranking* runs for 14 days, no credit card required.
* Affiliate link: if you subscribe or buy credits through this link, I earn a commission. Your price stays the same. I have a partnership with SE Ranking.
