Genjutsu earns 74/100 for making its video-to-video generator and one-time credit model easy to understand. Its most material weaknesses are the small mobile interaction layer, design-token drift, and limited self-serve documentation.
Reviewed by SiteList Engine · 12 of 13 dimensions · published Reviewed on October 24, 2026
Is this your site?
Claim itGenjutsu is a browser-based video-to-video generator and community template gallery for recreating viral motion with personal photos.
| Fact | Value |
|---|---|
| Domain | genjutsu.app |
| Category | AI video-to-video motion transfer for viral-video remixes |
| Pricing | One-time credit packs; 3, 7, or 11 credits per second at 480p, 720p, or 1080p |
| Pages crawled | 39 |
| Crawl date | 2026-10-24 |
| Platforms | Web |
Genjutsu presents a clear offer: a browser-based AI video generator that recreates viral motion with a user's photos. The strongest measured dimensions are accessibility at 100/100, usability at 95.2/100, and technical SEO at 85/100. The homepage and pricing routes also make the one-time credit model concrete, while the writing-quality score of 81.3/100 reflects practical copy and a usable three-step workflow on the flying-car page.
The review is held back by uneven execution. Design scores 54/100 because the mobile interaction layer is too small and the source CSS shows token drift. Performance scores 38/100, with low confidence: supplied PageSpeed Insights requests returned no lab or CrUX metrics, so there is no evidence-led LCP, INP, or CLS verdict. Editorial QA is 66/100, with the crawl described as a partial product, policy, template, and locale sample.
Genjutsu is legible within seconds as a video-to-video generator that puts a visitor's photos into the motion, camera, and timing of a reference clip. The hero, “Recreate Any Viral Video With Your Photos,” makes the mechanism concrete, while the screenshot places the motion-transfer workflow above the fold. Pricing adds exact rates of 3, 7, or 11 credits per second at 480p, 720p, or 1080p, plus one-time packs and automatic refunds. The gap is focus: pricing names weekly creators and high-volume teams, but the homepage does not choose one primary segment. Lead with weekly social creators, keep teams secondary, and add attributable before/after examples with the source clip, inputs, output, and resolution.
A prospect can classify Genjutsu quickly: the H1 and workflow describe an AI video generator that transfers a reference clip's motion, camera, and timing onto personal photos. Pricing answers cost and pack terms, while FAQ markup covers subscriptions, expiry, refunds, commercial use, and source-video limits. The unresolved decision is comparative. The site names Higgsfield and presents a different one-time-credit workflow, but the crawl found no comparison or alternatives page. Trust is similarly asymmetric: policy and operational answers are explicit, while the supplied artifacts show no attributable testimonial, case study, or review-count evidence. Add a comparison surface and one named creator example showing the source, inputs, output, settings, and result.
Genjutsu gives a first-time visitor a strong route from proposition to pricing, but the account-to-generation-to-purchase sequence is not explicit. The homepage presents “Add a reference video” and “Add your characters, products, or clothes,” while Pricing and Get credits both reach /pricing. Starter, Pro, and Max then show one-time prices of $9, $39, and $99 with credit counts and output estimates. The friction is competing vocabulary: Generate, Log in, Get credits, Buy credits, and Get Credits and Start are all plausible next actions. Make one primary start action account-aware, label login for returning users, and explain whether account creation happens before or after buying credits. The supplied mobile homepage shows no verified clipping blocker; mobile pricing parity was not captured.
The sampled pages show a strong semantic accessibility baseline, but the 100/100 score is provisional rather than a full conformance claim. Homepage, pricing, terms, privacy, refund, and templates samples expose header, navigation, main, and footer landmarks; heading sequences have no reported level skip; and the homepage input has a label or ARIA name. The crawl recorded zero missing-alt images across sampled pages, with one empty-alt image on the homepage and one on templates that may be decorative. Desktop and 390px mobile captures show no obvious horizontal clipping. Run Lighthouse or axe plus scripted keyboard and assistive-technology checks for focus order, trapping, accessible names, live regions, and media captions before treating the result as complete.
Genjutsu has a coherent light editorial direction, but its measured mobile controls and source-level design values need consolidation. The 390px homepage reported 19 small tap targets, including 32px header controls, 32px tabs, and a 40px search field; the input also uses a 14px computed font size. Increase actionable hit areas to at least 44×44px, give the input at least 16px text and 44px height, and preserve visual density through padding. The CSS extraction reports 82 colors, 39 font sizes, 33 radii, 25 spacing values, and seven font families, which is a drift signal rather than proof that every value is rendered. The homepage H1 is 16px against 16px body text, a 1.0× ratio. The 390px capture has scrollWidth 390 with no evidenced horizontal overflow.
The crawl cannot establish a Core Web Vitals verdict: every supplied PageSpeed Insights run returned no lab or CrUX metrics, so LCP, INP, CLS, and field TTFB remain unmeasured. The strongest measured risk is transfer volume. The homepage network summary reports 273,003 KB across 285 requests, including 270,418 KB from file.neomundo.ai across 197 requests. Templates and pricing report 2,340 KB and 2,406 KB with 33–34 blocking requests; the homepage reports 36. Trace the waterfall before assigning individual causes, inventory the file.neomundo.ai requests, remove duplicate or below-fold media, and establish a mobile transfer budget. The crawl also records 35 font-face rules. Subset fonts and defer noncritical gallery and application work, then rerun mobile and desktop measurements.
Genjutsu's strongest copy connects the visitor's action to the mechanism: choose or upload a viral clip, add photos, and transfer the motion, camera, and timing. Pricing supplies exact pack sizes, per-second rates, resolutions, and worked examples, while the flying-car template gives a clear three-step sequence from photo upload to download. The main editorial cost is repetition: pack cards, the credit table, FAQs, and later pricing sections repeat the same buying case. Keep one canonical explanation per fact and use the recovered space for pack selection. The flying-car page also makes a date-specific Tesla/Roadster claim without a supplied source link; add a primary source or remove the evergreen assertion. Treat the very high Japanese and Chinese sentence-length results as language-sensitive segmentation, not English clarity evidence.
Genjutsu's pricing page is a useful plan selector: Starter, Pro, and Max are separated by audience and volume, with prices of $9, $39, and $99, non-expiring credits, unit pricing, and output estimates at 480p, 720p, and 1080p. The page also states one-time payment, no subscription, and surfaces refund, privacy, storage, watermark, and commercial-use details. Pro carries a “Most popular” label and Max advertises the lowest price per credit, but neither cue explains the underlying tradeoff. Add one best-for sentence and one limitation to each plan, tied to monthly clip volume or resolution. Cards use 15-second estimates while the rate table uses a 5-second example; a shared duration-and-resolution calculator would remove that translation step. The English mobile capture preserves the comparison order.
Genjutsu's main observed growth risk is a public templates hub that returns 200 while declaring noindex, follow. The page is linked from homepage navigation and content, so the directive can suppress discovery of the community-template surface even though the route is publicly available. Remove noindex, retain a self-canonical, and verify indexability in Search Console after deployment; organic impact will require recrawling time. The wider stability picture is young rather than established: RDAP records registration on 2026-10-02 and Wayback has one snapshot dated 2026-10-03, so there is no credible two-year baseline. Other observed signals are sound: the nonexistent probe returns 404, HTTP root makes one 301 hop to HTTPS, and homepage raw and rendered word counts match at 1,712/1,712. Traffic, rankings, and demand remain unavailable.
Genjutsu clears the basic editorial bar on the sampled marketing surface: the homepage states the product, pricing gives usable numbers, and the flying-car template provides a concrete three-step path. Two checks need priority. First, the flying-car page states that Tesla showed its new Roadster at a SpaceX site in Texas on October 15, 2026, but the supplied prose has no primary-source link; source it, label it as context, or remove it. Second, homepage FAQ JSON-LD lists 8 questions while the complete heading tree exposes 4 FAQ question headings. Reconcile structured data with visible genuine content before relying on it. Pricing repetition is secondary. The crawl is partial at 39 of 40 pages, and the high Japanese and Chinese sentence-length metrics require language-aware segmentation.
Genjutsu's homepage FAQ answers practical questions about credit cost and expiry, supported uploads, refunds, commercial use, and privacy, giving evaluators useful policy context before generation. Turn the strongest FAQ answers into linked task guides with expected outcomes and recovery steps, add a Help entry, and publish a canonical index after the help center exists. The /llms.txt probe returned 404; that is best addressed after the underlying support content is available.
Genjutsu is broadly crawlable and server-rendered in the observed sample: robots.txt allows the public surface and declares a verified 64-URL sitemap, the nonexistent probe returns 404, and the HTTP root makes one 301 hop to HTTPS. The homepage exposes WebSite, Organization, WebApplication, and FAQPage JSON-LD, with raw and rendered word counts both at 1,712. The material blocker is /templates, which returns 200 but carries noindex, follow even though homepage navigation and content link to it. Remove that directive, retain a self-canonical, and verify coverage in Search Console. Hreflang is present for en, zh, ja, ko, es, fr, de, pt-br, and x-default on the homepage, but full reciprocity across all 64 sitemap URLs was not available. Core Web Vitals remain unjudged because PageSpeed Insights returned no usable metrics.
Genjutsu is a focused choice for social creators and high-volume content teams that want to remake reference videos with their own photos. The product is legible within seconds, pricing is based on one-time credit packs, and the homepage FAQ answers practical questions about cost, uploads, refunds, rights, and privacy.
The next improvements are concrete. Increase the size and usability of the mobile interaction layer, reduce design-token drift, and add public documentation beyond the homepage FAQ. Performance evidence also needs a clearer measurement path: the supplied PageSpeed requests produced no lab or CrUX metrics, leaving the 38/100 performance result low-confidence. The result is a credible, useful product surface with several fixable gaps before it feels fully mature.
This 13-dimension review uses the 39-page crawl captured on 2026-10-24, the public dimension score table, and product facts extracted from the current site, including JSON-LD FAQ data and observed page text. The review covers positioning, audience messaging, usability, accessibility, design execution, performance, writing quality, decision-support surfaces, risk and stability, editorial QA, docs and self-serve help, and technical SEO.
Dimension 13, Review-content integrity, is excluded as not applicable. Google Search Console access was not connected, so search and traffic enrichment remains pending. Performance should be read with care because the supplied PageSpeed Insights requests returned no lab or CrUX metrics. The crawl is also a partial product, policy, template, and locale sample. Read How SiteList scores for the scoring method.
Generation costs 3, 7, or 11 credits per second at 480p, 720p, or 1080p. A 5-second clip at 720p costs 35 credits, and the exact cost appears on the Generate button before submission.
No. Genjutsu uses one-time credit packs, so nothing renews and there is nothing to cancel.
Pack credits and credits returned through an automatic refund do not expire. The 50 free credits for new accounts are valid for 30 days.
Yes. Upload an MP4 or MOV file that is 4 to 30 seconds long, or start from a community remix.
Yes. Generated videos can be used in ads, stores, and social media subject to the terms of service. Upload only photos you have the right to use, and do not use someone else's likeness without permission.
Failed or timed-out generations are refunded automatically. A full refund is available within 7 days of purchase if no credits have been used; email support@genjutsu.app, and it is processed within 24 hours.