How long should an SEO slug be?

Three to five meaningful words, under about 60 characters. Long enough to be specific and unique, short enough that search results do not truncate it.

Hyphens or underscores?

Hyphens. Google documents hyphens as word separators; underscores join words together, so file_name reads as one token rather than two.

Should the slug match the title exactly?

It should match the intent, not the words. Titles are written for click-through and often carry filler that adds nothing to a URL — trim it.

Do keywords in the URL still help rankings?

Marginally, as a direct factor. The larger effect is on click-through rate: a readable, relevant URL in the result gets clicked more than a numeric one, and engagement is not marginal.

SEO Slug Generator

What Search Engines Do With Slugs

Google has said the URL is a small ranking factor. That understates its practical value: the slug appears in the search result, in shared links, and in every citation of the page. It is one of the few signals that is simultaneously read by crawlers and by humans deciding whether to click.

Length

Search results truncate the displayed URL at roughly 60–70 characters including the domain. Longer slugs are not penalised, but the part a searcher reads gets cut off.

Title	Slug	Length
How to Build a REST API in Node.js in 2026	`how-to-build-a-rest-api-in-nodejs-in-2026`	41
Stop words removed	`build-rest-api-nodejs-2026`	26
Over-trimmed	`rest-api`	8 — too generic to rank

The middle row is the target: specific enough to be unique, short enough to display.

Stop Words

Removing a, the, of, in and to shortens URLs without losing meaning — most of the time. Keep them when the phrase inverts without them: stripping "not" from why-not-to-use-orm produces the opposite claim.

Dates and IDs

Pattern	Problem
`/2024/01/my-post`	Visibly ages; forces a redirect when you refresh the content
`/post/48211`	Meaningless to readers and crawlers alike
`/blog/p=123`	Query parameters are crawled less predictably than paths
`/how-to-build-a-rest-api`	Readable, stable, descriptive

Changing a Slug Later

Every inbound link, every ranking signal and every bookmark points at the old URL. If you must change one:

1. Serve a 301 from the old slug to the new one — permanent, not 302. 2. Keep the redirect forever; there is no safe point to remove it. 3. Update internal links to point directly at the new URL, so crawlers do not hop. 4. Never chain redirects — each hop loses a little and adds latency.

The cheapest fix is to get the slug right before publishing, which is what this tool is for.

Slugs in a Real Content System

Decision	Recommendation
Generate from	The title, at draft time, then let the author edit
Freeze on	First publish — after that, changing it costs a redirect
Collisions	Append `-2`, not a timestamp or a random suffix
Depth	`/blog/post-name`, not `/blog/2026/08/category/post-name`
Case	Lowercase, always — URLs are case-sensitive on most servers

Store the slug as a column with a unique index rather than deriving it on read. Deriving it means a title edit silently changes every link to the page.

Redirects When One Changes

``301 /old-slug → /new-slug`

Permanent, kept forever, and never chained. Two hops lose more than one, and a chain of three is treated by crawlers as a soft failure. Keep a redirect table rather than editing the previous rule in place.

`Non-Latin Titles`

Transliteration is the pragmatic choice for a Latin-script audience — Приветbecomesprivet. If your readers use the script, keep it: modern browsers and search engines handle IDN and percent-encoded UTF-8 paths correctly, and/статья` is more readable to a Russian speaker than any transliteration. The wrong answer is silently dropping the characters, which produces an empty or mangled slug.