How Rankevra works
Rankevra crawls a site you own, audits what it finds, writes articles for the gaps, and publishes them where the site lives. Everything below is generated from the same configuration the product runs on.
Crawling
A crawl fetches your pages and records what is actually there — titles, headings, meta descriptions, internal links, images and their alt text. It reads the rendered page, so a site built with JavaScript is measured as a visitor sees it rather than as the raw HTML.
How many pages a single crawl covers, and how often you may run one, are set by your plan — see the table below. Only sites you own should be crawled; the terms are explicit about that.
What each plan allows
| Free | Starter | Pro | Agency | |
|---|---|---|---|---|
| Sites | 1 | 1 | 3 | 25 |
| Crawls | 1 per month | 5 per day | 5 per day | 5 per day |
| Pages per crawl | 50 | 500 | 5,000 | 10,000 |
| Articles per month | — | 30 | 90 | 400 |
| Article refreshes per month | — | 8 | 25 | 150 |
| Team members | 1 | 1 | 3 | Unlimited |
Articles and refreshes are separate pools on purpose: regenerating an existing article never eats the allowance you are paying for to write new ones. A dash means the plan has no AI content at all.
One Evra subscription covers Rankevra alongside every other product — see pricing.
What needs which plan
| Feature | From | What it is |
|---|---|---|
| AI articles | Starter | Draft full articles from a topic or a keyword, grounded in the business details on the site. Includes the content calendar and Autopilot, which schedules and writes them without you asking each time. |
| Keyword research | Starter | Real search volume, difficulty and intent for the keywords a site could rank for, rather than a guess derived from the page text. |
| Internal linking | Starter | Finds pages on the site that should link to each other and proposes the link with its anchor text, so new articles are not published as orphans. |
| Publishing | Starter | Sends a finished article to where it should live — WordPress, WordPress.com, a webhook, or an embed widget — instead of leaving it as a draft to copy out by hand. |
| Search Console | Pro | Connects Google Search Console so impressions, clicks and position come from what Google actually recorded, not an estimate. It is what turns 'this article exists' into 'this article is ranking'. |
Articles
An article is drafted from a topic or a keyword and grounded in the business details on the site, so it argues from what you actually do rather than from the topic in the abstract. You choose how deep it goes:
- Standard — A focused article on one topic. The default, and the right choice for most posts.
- Detailed — Longer, with more sections and more supporting detail. Costs more of the monthly article allowance.
- Authoritative — The long-form treatment, for the pages meant to be the definitive answer. The most expensive per article.
Cover images
Cover art is opt-in. New sites generate no image and are billed for none; a site turns it on by choosing one of 11 styles in its settings, and NONE means articles publish text-only.
The reason it is opt-in rather than on: an AI cover costs roughly twice what the words cost. Charging every site for a picture it never asked for is most of the price of an article spent on the part nobody requested. Existing sites keep whatever they already chose.
Autopilot
Autopilot writes and schedules articles on its own, at a frequency and depth you set per site, inside the same monthly allowance as manual articles — it does not have a separate budget, so it can never spend more than the plan you are on.
Publishing
A finished article can be sent straight to where the site lives:
- WordPress (self-hosted) — Your own WordPress via its REST API and an Application Password.
- WordPress.com — Authorised over OAuth — no password stored, and you pick which of your sites receives the post.
Credentials are encrypted at rest and never shown back to you after saving. With no publish target connected, articles stay as drafts here rather than silently going nowhere.
Measuring whether it worked
Connect Google Search Console and the numbers come from what Google recorded — impressions, clicks and average position per page — instead of an estimate derived from the page itself. It is the difference between knowing an article exists and knowing it ranks.
Something here wrong or missing? Tell us — this page is generated from the same configuration the product runs on, so if it disagrees with what you see, that is a bug worth hearing about.