What score should I aim for?

Aim for 65 or above in semantic mode. Scores between 45 and 65 are improvable with stronger topical vocabulary and more direct answers to the query.

Does a high score guarantee ranking?

No. Relevance is one signal. Authority, backlinks, UX, and intent match all matter. A high score means the content is topically aligned — not that it will automatically rank.

How is the heuristic fallback computed?

Jaccard similarity on non-stop word tokens between the query and the passage. It only captures shared vocabulary, not conceptual proximity.

Can I fetch a URL instead of pasting text?

Yes. Use the Fetch a page input to pull content from any public URL. The main body text is extracted and used as the passage.

What passage length gives the best results?

At least 100 words gives the embedding enough signal. For a full article, paste the introduction or the section most likely to rank — not the whole page.

Content & Writing

New

Live

Relevance Scorer

Score how relevant a passage is to a query using embedding cosine similarity.

Query

The search query or topic you want to match against.

Passage

Paste text or fetch a URL. The passage is scored against your query.

Fetch a page

Semantic assist — embeds query and passage via /api/embeddings and scores cosine similarity. Disable to use keyword-overlap fallback only.

Relevance score

Low relevance

Keyword overlap

Tip: The passage covers a different topic than the query. Rewrite with the query intent as your primary goal.

Keyword overlap: 4 — counts shared non-stop words between query and passage. Enable Semantic assist for embedding-based scoring.

X / Twitter LinkedIn Facebook

Start here · Why does query–passage relevance matter for SEO?

Search engines measure whether a page actually answers the query behind a keyword. A page that mentions the keyword but drifts into tangential content scores lower on topical relevance.

This tool embeds your query and your passage independently and scores cosine similarity between them. A score of 70+ means strong conceptual alignment. Below 45 means the passage needs a tighter topical focus.

When Semantic assist is disabled, keyword-overlap Jaccard similarity serves as a fast, local approximation — weaker but still useful for quick checks.

When to use this tool

You have a real task today
Use Relevance Scorer when you need to score how relevant a passage is to a query using embedding cosine similarity. That might be a launch checklist, a client question, or a ticket that says "quick check before we publish."
You are learning what the numbers or text mean
Run Examples below with a tiny input. Run one real draft through the tool before you rewrite a whole content library. Watching one clean input change is the fastest way to build intuition.
You are scaling across many URLs or writers
Prove the workflow once here, then repeat the same pattern in a spreadsheet, Notion database, or your SEO platform. This page is the sandbox; your tracker is the production line.

Examples

Walk through these with the form above — they are practice scenarios, not live data.

Minimal first run (recommended)

Try this

Give the smallest honest input the labels allow — one short paragraph, one keyword, one staging URL, etc. Avoid pasting an entire site on the first try.

What to look for

Notice the shape of the result: lists, scores, generated copy, or checks. Relevance Scorer makes sense once you have seen one full input→output cycle.

Real page or real draft

Try this

Paste a draft, outline, keyword, or page text from one URL. Use something you would actually ship or review, not placeholder text.

What to look for

Look for the recommendations that make the page clearer, more complete, or easier to scan. If the result surprises you, change only one variable and run again.

Short tutorial

Follow in order the first time you use the tool; later you can skip to the step you need.

Step 1 — Read labels and placeholders
The Relevance Scorer form tells you whether to fetch a https:// URL, paste raw text, or drop HTML. Placeholders show the format the tool expects.
Step 2 — Enter the smallest useful input
Start with the minimum that still describes the page or problem. Add more only if the output is empty, thin, or obviously wrong.
Step 3 — Run and scan the output
Use the main action (Analyze, Fetch, Generate, Score, etc.). Read headings and numbers before you copy anything — you are checking for sanity, not chasing perfection on the first pass.
Step 4 — Copy, file a ticket, or iterate
Paste the snippets you need into your CMS or doc. If something is off, change one field at a time so you learn what moved the result.

More detail

Relevance scoring is what separates content optimization from keyword stuffing. You can repeat a keyword ten times and still have a passage that fails to answer the searcher's actual question. Embedding-based relevance catches that gap: if your passage is conceptually distant from the query, you will see it in the score before you publish. Use this tool when writing a new section, evaluating a draft against a target keyword, or checking whether a repurposed article still aligns with its intended query.

FAQ

What score should I aim for?: Aim for 65 or above in semantic mode. Scores between 45 and 65 are improvable with stronger topical vocabulary and more direct answers to the query.
Does a high score guarantee ranking?: No. Relevance is one signal. Authority, backlinks, UX, and intent match all matter. A high score means the content is topically aligned — not that it will automatically rank.
How is the heuristic fallback computed?: Jaccard similarity on non-stop word tokens between the query and the passage. It only captures shared vocabulary, not conceptual proximity.
Can I fetch a URL instead of pasting text?: Yes. Use the Fetch a page input to pull content from any public URL. The main body text is extracted and used as the passage.
What passage length gives the best results?: At least 100 words gives the embedding enough signal. For a full article, paste the introduction or the section most likely to rank — not the whole page.