SMF·UBERSUGGEST
Can a prompt replace Ubersuggest?
SEO & marketing — keyword research, backlinks and rank tracking
Exhibit tracking slip
Verdict
Ubersuggest's numbers come from licensed clickstream and index data, and there is no version of a personal build that produces them. What you can build from entirely legitimate public sources is the content-gap half: crawl the sitemaps of sites you compete with, extract what they have written about, cluster it, and see what is missing from your own. That answers a real question Ubersuggest is used for, without a single fabricated volume figure.
Exhibit A — The prompt
Received on31.07.2026Build a content gap analyser that works entirely from public sitemaps and page content. No search volume, no traffic estimates, no scraped SERPs.
Input: a list of domains — the user's own plus a handful of competitors. For each, discover the sitemap from robots.txt or the conventional locations, and parse it including sitemap index files.
Crawl: fetch each URL politely with a descriptive user agent, honour robots.txt and a rate limit, and cap per-domain URLs. Extract title, meta description, H1 and H2 headings, the first paragraph, the visible word count, and the publication and modified dates where present in structured data.
Topic extraction: from each page's title and headings, extract candidate topic phrases (noun phrases, with stopwords and boilerplate removed). Keep the raw text alongside so any topic can be traced back to the page it came from.
Clustering: embed each page's title-plus-headings text locally and cluster across all domains at once, so the same topic covered by three different sites lands in one cluster regardless of wording. Label each cluster with the most representative phrase from its members.
The report, which is the point:
- Gaps: clusters where competitors have pages and the user's domain has none, ordered by how many competitors cover it — a topic three of them wrote about is a stronger signal than one.
- Thin coverage: clusters where the user has a page but it is markedly shorter than the competing pages in the same cluster.
- Overlap: clusters where the user has two or more pages, which is where internal competition usually hides.
- Freshness: clusters where competitors' pages are substantially newer than the user's.
Every figure in every view must be something the tool measured on a page it fetched: word counts, dates, page counts. Never estimate a search volume, a difficulty or a traffic number, and say in the README that the absence of those figures is deliberate and is the real limitation against a paid tool.
Storage: SQLite, so a crawl can be re-run and compared against a previous one.
Out of scope: search-engine result scraping, backlink data, keyword volume from any source, and rank tracking.
Opening prefills the prompt — press enter to run it.
Exhibit B — What you lose
- B.1 search volume, CPC and difficulty for every keyword
- B.2 traffic estimates for any domain
- B.3 the backlink data behind the link-opportunity reports
- B.4 the browser extension that overlays metrics on a live search page
Exhibit C — Why people still pay: proprietary web index and data acquisition
Because a topic gap tells you what to write and a volume figure tells you whether it is worth writing, and only one of those is available without buying a dataset.
Questions
Can I import anything from Ubersuggest?
Its keyword lists export as CSV and can be pasted in as notes, but nothing in this build consumes volume or difficulty figures, so they would sit unused. The useful migration is just the list of competitor domains.
How do I decide what to write without volume data?
By coverage weight and by your own Search Console data, which is free and genuinely yours. A topic three competitors cover and you do not is a strong signal on its own; pair this tool with a Search Console archive and you have most of what a keyword tool provides for a site you already run.
What does it cost to run?
Nothing but time and a small VPS if you want it scheduled. Crawling public sitemaps is free, embeddings run locally, and the whole dataset for a handful of competitor sites fits comfortably in SQLite.
What is the one thing that does not survive the rebuild?
Knowing the size of the prize. Ubersuggest tells you a topic gets 8,000 searches a month; this tells you three competitors wrote about it. The second is directionally useful and the first is what you actually want when deciding where to spend a week.
Related tools
Receipt