Jev SEO

Guide · Tools

Two tools called jev-seo, and which one you want

Search for "jev-seo" and you get two different open-source projects with the same name, both released in September 2026, plus a pile of articles about a model called Jev. One is a Python audit of a live website that runs as a Claude Code skill. The other is a Rust command-line tool that audits files on your disk. They don't share code or authors. Here is how to tell them apart in a minute.

Checked against both projects' documentation on .

Why there are two

Both tools are built on Jev, the decision model TypeSafe AI released on 15 September 2026. Jev doesn't write text; it answers narrow questions with a typed answer and a probability, which makes it a good fit for SEO judgments such as "is this page helpful for this search" or "how likely is an AI answer to cite this". Within a week, several people had built SEO tools on it and two of them picked the obvious name. A third thing you will meet in the same results, "Jev for SEO", is usually an article about using the model itself, not a tool at all.

At a glance

jev-seo by Daniel Agricijev-seo by Akash Priyadarshi
What it auditsA live website, crawled from its homepage URLHTML and Markdown files on your machine, plus live search results
Written inPythonRust, one binary
How you run it/jev-seo https://example.com in Claude Code, or bin/jevseo run from a terminalcargo install jev-seo, then sixteen commands; also an MCP server for coding agents
Checks52 rules tied to Google Search Central, Core Web Vitals from PageSpeed Insights, and a Jev judgment of every page: intent, helpfulness, trust, quotability for AI answersBatch on-page audit (thin pages, title collisions, orphans, AI vocabulary), Schema.org validation, AI crawler permissions in robots.txt, a 1 to 10 GEO citation score, rank drift in a local SQLite file
Search dataOptional: rankings, keywords, competitors and backlinks from DataForSEO, about 0.30 USD a siteFree DuckDuckGo endpoints for SERP radar, keyword autocomplete and briefs; no paid data API
What you getA designed PDF, an Excel action tracker and a Markdown report, all from one audit fileTerminal output and Markdown briefs
CostFree; the content review costs about one US cent per 60 pages in Jev usageFree; a free TypeSafe key is optional for the semantic scoring
LicenceMITMIT
Wheregithub.com/AgriciDaniel/jev-seojevseo.vercel.app

Facts come from each project's README and website as of . The Rust project's GitHub address linked from its site was not reachable when we checked; its website and Product Hunt page were.

The Python one: an audit report for a whole site

Daniel Agrici's jev-seo takes one homepage URL and does what an agency audit does: it crawls the site politely (robots.txt, sitemaps, redirects, JavaScript-only pages rendered in a real browser), runs 52 checks that each cite a Google Search Central page, measures Core Web Vitals, and asks Jev a fixed set of questions about every page. Code scores and ranks every fix; the model only judges meaning. The result is a PDF you can send to a client, an Excel sheet to track the work and a Markdown file for your docs. With a DataForSEO account it adds rankings, keywords within reach, competitors and backlinks.

It is built for Claude Code: type /jev-seo and a URL, and Claude runs the audit in the background, checks the surprising findings against their evidence and writes the executive summary. Our step-by-step guide is how to use Claude for SEO audits, and the example report shows the finished output.

The Rust one: a fast local checker and search radar

Akash Priyadarshi's jev-seo is a single Rust binary that audits the files of a site you have on disk, hundreds of them in under a second: thin pages, duplicate titles, orphan pages, "AI vocabulary", Schema.org validity, sitemap validity and which AI crawlers your robots.txt lets in. It also scores a page's chance of being cited in AI answers, tracks rank drift over time in a local SQLite file, pulls live competitor results and autocomplete keywords from DuckDuckGo, and turns a SERP into a writing brief. An MCP server exposes all of it to Claude Code, Gemini CLI and similar agents. It positions itself as a zero-cost alternative to a Semrush or Ahrefs seat, and it does not render pages, measure Core Web Vitals or produce a client-ready report.

Which one do you need?

  • You want a report on a live website, to act on or to send to someone. The Python one. It is the only one of the two that crawls a site from its URL and writes a PDF.
  • You keep a site as Markdown or HTML in a repository and want checks before you publish. The Rust one. It runs in a second on local files and fits a pre-commit hook or a CI step.
  • You want rankings, backlinks and competitors. The Python one, with a DataForSEO account; it is paid data, about 0.30 USD a site. The Rust tool deliberately avoids paid data APIs and uses DuckDuckGo instead, which gives you competitor pages and keyword ideas but not ranking positions or backlink counts.
  • You want AI-search checks. Both. The Python one rates every page's quotability and checks AI crawler access as part of its audit; the Rust one has a dedicated GEO citation score and a crawler radar command.
  • You don't want to install anything. Neither, then: Jev SEO, this site, runs the Python engine in the browser. Paste an address and the report is ready in minutes, with the search data included on paid checks.

Plenty of people would use both: the Rust tool while writing, the Python audit when a site is live.

Where this site fits

Jev SEO (jev-seo.com) is a web service built on Daniel Agrici's open-source engine. It is not affiliated with the Rust project or with TypeSafe. We run the same crawl, the same 52 checks and the same per-page review, add the search data, and hand you the PDF, Excel and Markdown without Python, API keys or a terminal. If you'd rather run the engine yourself, the GitHub link above is the place, and nothing on this site is hidden from it.

Common questions

Are the two projects related?

No. Different authors, different languages, no shared code. They share a name because both are built on Jev and both do SEO.

Is "Jev SEO" the same as "jev-seo"?

"Jev SEO" is the name of this web service. "jev-seo" is the name both open-source tools use. When people write "Jev SEO" in a video or a post from September 2026, they usually mean the Python skill.

What about "Jev for SEO" articles?

Those are about the model: how to use TypeSafe's Jev directly for keyword classification, internal linking and similar jobs. Useful if you build your own tooling; not a tool you can install.

How do I search for the right one?

Add a word. "jev-seo claude code" or "jev-seo pdf report" finds the Python audit; "jev-seo rust" or "jev-seo cargo" finds the Rust tool.

Comparing more options? The best Claude SEO tools covers the other skills and MCP servers.

The site audit, with nothing to install

Start free: 1 free check, then $0.99 per check.