Everygen AI Review: clear offer, costly mobile path (68/100)

Field	Value
Domain	everygenai.org
Category	AI text-to-video and image-to-video generation
Pricing	Paid; credit-based
Pages crawled	10
Crawl date	2026-10-11

Everygen AI makes its category and core workflow easy to understand: visitors can move from text or images to AI video, choose models, and control outputs. Its generator-first homepage and examples support a strong first impression, while pricing is transparent about credits, plans, model rates, and image/video allowances. The experience loses ground in execution. Usability remains fair because mobile startup is slow, plan comparison is demanding, and an empty blog section feels unfinished. Performance is the clearest constraint. Design execution also scored weak, with 37 small mobile tap targets and confirmed contrast audit failures.

Everygen AI states its job clearly in five seconds: text and images to video, with social clips, product demos, ad visuals, and controls for model, resolution, duration, and framing. The homepage shows 17 examples. Name one primary buyer and add proof tied to that workflow.

The homepage puts Prompt, Text to Video, Image to Video, model, aspect ratio, duration, resolution, framing, and Generate before explanatory copy. Pricing then combines packs, subscriptions, shared balances, model rates, allowances, and commercial-use differences. Preserve the generator-first path and add a representative workflow cost or job-based guide.

Visitors understand the offer and reach pricing, but mobile startup and comparison effort keep usability fair. Mobile measured 16.95 s LCP, 22.19 s Time to Interactive, and 48,173 KiB. Pricing uses generic CTAs, while /blog says “No posts yet — check back soon” and is orphaned. Prioritize the generator shell, label plans, and publish an initial blog post or remove the public blog entry until it has content.

Landmarks and generator labels are present, and no positive tabindex values were found. The homepage has 1 unlabeled input out of 2; Lighthouse reports ARIA and label-name failures, plus contrast failures on the homepage and video generator. No skip link was found. Label the input, correct ARIA, add a keyboard-visible skip link, and retest contrast.

The 390px render has no horizontal scroll and the homepage has orderly headings, but the audit recorded 37 small tap targets, including 36×36px header buttons and 26px or 30px generator controls. CSS extraction found 62 colors, 21 radii, and 19 font sizes. Raise controls to 44×44px, repair contrast, and consolidate tokens.

Mobile delivery is the clearest constraint: homepage LCP was 16,952 ms, FCP 4,201 ms, Speed Index 5,365 ms, and TTI 22,191 ms, with a 48,173 KiB payload. Estimated image savings are 4,041 KiB and unused JavaScript 298 KiB. Right-size above-fold media, defer below-fold assets, split route code, and re-test. These are lab measurements.

The hero is direct: “Create social clips, product demos, and ad visuals from text or an image.” Pricing gives “10 credits $9.90 ≈ 5 videos.” Polish drops where the blog is empty and tool pages repeat the same three steps. Keep the plain language, remove the empty public promise, and tailor instructions to each modality.

Pricing supplies USD prices, credit quantities, approximate output counts, annual equivalents, and model-rate tables; it also explains the shared image/video balance. One-time packs and annual subscriptions share a long surface with generic buttons. The tables contain 4 image columns and 20 video rows. Separate models, label actions by plan, and add filters.

The crawl supports factual reporting of prices, credits, billing, and model rates, not claims about comparative quality or hands-on performance. The candidate inventory contains 1 pricing surface, 0 review-content candidates, and no freshness dates. Any future comparison should add dated methodology and first-party test evidence.

RDAP dates registration to 2026-10-08, three days before the crawl; Wayback has 1 snapshot and no CrUX field metrics are available. Pricing reaches 646px at a 390px viewport. Monitor discovery and fix overflow.

Compliance copy is thorough, but QA is uneven. The homepage has 25 showcase images without descriptive alt attributes and pricing has 8 more. FAQ wording “Nano Banana Pro (Nano Banana 2)” differs from generation-form names. Add descriptive alternatives where informative and standardize model names across FAQs, forms, and catalog routes.

Sampled pages return 200 OK, canonicals self-reference, core text is server-rendered, HTTP redirects to HTTPS, and robots.txt points to a sitemap listing 11 pages. The main gaps are /pricing overflow to 646px at a 390px viewport and a 720 KB logo rendered at 24px to 32px. Wrap tables responsively and optimize the logo.

Everygen AI is a credible fit for creators and marketing or product teams that want a direct path to AI-generated social clips, product demos, and ad visuals. The product surface communicates its core workflow clearly, and the pricing page provides useful detail about credits, prices, billing, and model rates. The next improvements are concrete: reduce the homepage mobile delivery burden, enlarge the 37 small mobile tap targets, address confirmed contrast failures, and make the two purchase models easier to compare.

This 13-dimension review uses crawl evidence from 10 pages captured on 2026-10-11, alongside the public dimension score table and site facts. GSC was not connected, and no CrUX field data was supplied. See How SiteList scores for the review method.