Skip to content
SiteList

Chraft Review: strong core, slow delivery (71/100)

Chraft scores 71/100, with clear positioning, strong usability, and consistent design for content creators on TikTok, YouTube, and Instagram. Its most material weaknesses are a 10.1-second mobile LCP, dense decision-support surfaces, and editorial issues including cloneviral.ai canonicals and duplicate titles.

Reviewed by SiteList Engine · 11 of 13 dimensions · published Reviewed on September 4, 2026

Is this your site?

Claim it

Quick facts

Fact Value
Domain chraft.ai
Category AI video generation for social media
Audience Content creators for TikTok, YouTube, and Instagram
Pricing Unknown
Pages crawled 39
Crawl date 2026-08-31

Executive summary

Chraft scores 71/100, led by clear positioning at 87/100, usability at 85/100, design execution at 87/100, and technical SEO at 82/100. The site identifies content creators as its audience and presents AI video creation without manual editing as the core proposition.

Performance is the largest constraint. The homepage records a 10.1-second mobile LCP, a 19ms TTFB, and a 76MB payload. Decision-support surfaces score 65/100 because the pricing and comparison grids provide limited guidance across five plans and three decision axes.

Writing quality scores 58/100, while Editorial QA scores 52/100. The supplied findings identify CloneViral canonicals, duplicate titles across four routes, loading-state pages, and cliché-heavy informational copy. Review-content integrity is stronger at 75/100 but still needs methodology and tradeoff detail in recommendations.

Evidence
Overall score
71/100
Mobile LCP
10.1 seconds
Homepage payload
76MB

01 · First impressions & positioning — a specific conversational video promise

Chraft's positioning is clear: “Chat to Create Viral AI Videos” promises conversational creation without manual editing. The site reinforces that promise with named testimonials, 50K+ videos generated, and detailed feature lists. Recognition is weaker because the brand name and chraft.ai domain do not align cleanly across surfaces. The “AI-native” category is also less clear than a conventional description, and jargon density reaches 3.5 occurrences per 1,000 words for terms such as “viral,” “cutting-edge,” and “revolutionize.” Use a conventional category such as “AI Video Generation Platform,” then align schema and metadata and replace broad claims with concrete functionality.

Evidence
Positioning statement
Chat to Create Viral AI Videos
Jargon density
3.5/1,000 words

02 · Audience & messaging — creators are named, but the hero stays implicit

Chraft targets content creators working across TikTok, YouTube, and Instagram, with a promise of viral video creation without manual editing. Named testimonials and customer counts add proof, but the audience is not stated directly as a segment in the hero. The copy also leans on terms such as “AI-native” and “revolutionize,” which are less useful to creators than plain descriptions of the workflow. Make the audience explicit in the opening copy, simplify the terminology, and add specific outcomes to testimonials. The supplied evidence cites follower growth from 50K to 500K, but the testimonials do not consistently explain the result or its context.

Evidence
Named channels
TikTok, YouTube, Instagram
Customer proof
50K+ videos generated

03 · Usability — clear homepage paths, with conversion actions below the fold

Chraft's homepage communicates its proposition with a bold headline, video examples, and a visible “Start for Free Now” action. The main usability cost appears on pricing and mobile: the pricing page is two clicks from Home, its conversion button requires scrolling, and mobile navigation does not surface Pricing or Start Trial immediately. Documentation and blog navigation are organized, but the homepage lacks direct links to core product features. Move the primary pricing CTA above the fold, add a contextual “Choose Your Plan” heading or breadcrumb, prioritize high-value actions in the mobile menu, and standardize link styling across sections.

Evidence
Pricing depth
2 clicks: Home → Pricing
Primary CTA
Start for Free Now

04 · Design execution — strong token discipline, but contrast remains 2.8:1

Chraft's design system is consistent and modern, with CSS custom properties covering colors, fonts, and spacing. Mobile checks found no horizontal scroll or tap-target failures. The material defect is contrast: light gray #9CA3AF on white produces a 2.8:1 ratio in the hero and pricing section, below the 4.5:1 WCAG AA requirement for body text. Pricing cards also vary in padding and font size. Darken the affected text to a compliant shade, then standardize card padding and type tokens so the visual system remains consistent as plan content changes.

Evidence
Contrast ratio
2.8:1
Required ratio
4.5:1

05 · Performance — 10.1 s mobile LCP driven by a 76 MB payload

Chraft's homepage is critically slow on mobile: LCP is 10.1 seconds, despite a fast 19ms TTFB, and total payload weight is 76.3MB on mobile and 76.7MB on desktop. The LCP hero image is 7.2MB, naturally 3,168px wide but rendered at 330px, with no preload or high-priority hint. The network log also records 49 blocking requests, including synchronous analytics. Resize and compress the hero asset, preload it with high priority, lazy-load non-critical images, and defer third-party scripts. Font loading needs the same treatment to prevent FOUT.

Evidence
Mobile LCP
10.1s
Mobile payload
76.3MB
Blocking requests
49

06 · Writing quality — clear product story weakened by legacy references

Chraft's homepage explains its AI video creation proposition, but editorial polish drops on supporting pages. The About page uses cliché-heavy phrases including “Advanced AI Video Generation Technology,” “empower anyone, anywhere,” and “Modern AI Technology.” The /mcp-skills canonical points to cloneviral.ai, and CloneViral asset references create a legacy-brand mismatch. Pricing, mcp-skills, and characters also lack H1 headings, while comparison pages average 32.8 to 34.1 words per sentence. Replace slogans with concrete capabilities, correct the legacy references, add one specific H1 per page, and break long comparison sentences into direct points.

Evidence
Comparison sentence length
32.8–34.1 words
Legacy canonical
https://www.cloneviral.ai/mcp-skills

07 · Decision-support surfaces — five plans collapse into a 60-cell feature grid

Chraft's pricing surface presents five plans across 12 feature rows, creating 60 cells without enough guidance for choosing among them. The table offers only three decision axes—price, credits, and support—and comparison tables include three all-checkmark rows that provide no meaningful distinction. Replace feature-list scanning with a segmented recommendation layer that states who each plan suits and what each tradeoff costs. Strengthen the comparison axes around model access, duration, and resolution, and replace identical checkmarks with informative differences. The goal is a decision path, not a larger inventory of features.

Evidence
Plan options
5
Feature grid
12 rows / 60 cells

08 · Review-content integrity — honest comparisons need method and tradeoffs

Chraft's comparison content is directionally honest, but its recommendations do not show enough of the reasoning behind them. The “Quick Verdict” says “Chraft is Better for Most Users” without naming a tradeoff, and the Chraft versus Runway ML page lacks a methodology block. Add the comparison criteria and method directly beside the verdict, then state where the alternative is stronger—for example, advanced video editing tools. Each comparison page should also carry an affiliate disclosure where relevant. These additions make the recommendation auditable without changing its conclusion.

Evidence
Verdict needing tradeoff
Chraft is Better for Most Users
Missing method
No methodology block on Chraft vs Runway ML

09 · Risk & stability — canonical drift and duplicate metadata threaten discoverability

Chraft has a sound indexability base, but several correctable signals could erode search visibility as the domain matures. The domain is under seven months old, /mcp-skills points its canonical to cloneviral.ai, and the homepage title is duplicated across four pages. The homepage hero is 7,078KB at a 330px display width, and six pages contain images missing alt text. Fix self-referencing canonicals and unique metadata first, then compress the hero and add alt text. Connect Google Search Console to monitor index coverage and click trends as these changes take effect.

Evidence
Domain age
<7 months
Duplicate homepage title
4 pages

10 · Editorial QA of content — live routes expose unfinished publishing controls

Chraft's publishing hygiene is inconsistent across important routes. The /mcp-skills canonical points to cloneviral.ai, /evostudio exposes a seven-word “Loading” shell, and four distinct URLs share the title “Chat to Create Viral AI Videos | Chraft.” Long-form pages also contain multiple H1 tags. These are visible QA failures rather than abstract editorial concerns: readers and search engines receive incomplete or ambiguous page signals. Correct the canonical, complete or gate the loading route, assign unique titles, and enforce exactly one H1 per page before publishing further template content.

Evidence
Loading shell content
7 words
Duplicate title scope
4 URLs

11 · Technical SEO — crawlable foundations offset by canonical and metadata debt

Chraft's technical foundation is strong: Googlebot is permitted across high-value paths, sitemaps are valid, and sampled pages render full content without client-side dependency risks. Canonical hygiene is the main exception. /evostudio, /docs/evostudio, and /comparison point to the homepage, while /mcp-skills points to cloneviral.ai. The homepage title is duplicated across four pages, and /blog and its tutorial view share metadata. Fix canonicals and write unique titles and descriptions, then compress the 7,078KB hero, add missing alt text, and reduce the HTTP root's two-hop redirect chain to one 301.

Evidence
Hero banner
7,078KB displayed at 330px
HTTP redirect chain
2 hops

Verdict — 71/100: strong core with costly delivery and editorial gaps

Chraft is a credible fit for social-video content creators who value clear positioning and a consistent product presentation. Its strongest evidence is the combination of 87/100 positioning, 85/100 usability, and 87/100 design execution.

The priority fixes are concrete: reduce the 76MB homepage payload and 10.1-second mobile LCP; repair canonical and duplicate-title defects; and add recommendations, tradeoffs, and methodology to pricing and comparison surfaces. Chraft has a sound core, but these issues need attention before the site can support a polished customer-facing experience.

Evidence
Overall score
71/100
Positioning
87/100
Usability
85/100

Methodology & data notes

This 13-dimension review uses crawl evidence from 39 pages crawled on 2026-08-31, together with the public dimension score table. It presents a 13-dimension review while the supplied frame contains the scored dimensions used for this article, including the two dimensions identified as excluded where applicable.

Google Search Console access was unavailable. See How SiteList scores for the review method and data notes.

Questions buyers actually ask

Who is Chraft for?

Chraft targets content creators making videos for TikTok, YouTube, and Instagram, especially those seeking AI generation without manual editing.

What is Chraft’s biggest weakness?

Performance is the largest weakness. Mobile LCP is 10.1 seconds, while the homepage payload is 76MB and TTFB is 19ms.

How usable is Chraft?

Usability scores 85/100. The homepage communicates its value with clear calls to action, though pricing conversion elements need better placement.

How does Chraft compare with other tools?

Chraft publishes comparison content, including Chraft versus Runway ML, but its recommendations need explicit tradeoffs and methodology.

How was Chraft reviewed?

SiteList reviewed 39 pages as a 13-dimension review, using crawl evidence and the public dimension score table.

How this review was made

SiteList reviewed chraft.ai on September 4, 2026 — pages, screenshots, performance runs, structured data and public records — then scored it across 11 of the 13 public dimensions. Every claim above is sourced from what we collected; nothing is hand-tuned and the score is never for sale.

Not covered in this write-up: Accessibility (not applicable), Docs & self-serve help (not assessed). Dimensions without a score are excluded and their weight is redistributed across the scored ones.

Pending enrichment (data we could not fetch this run): readability_api, plagiarism_check, fact_checking_api

Read the full methodology

71/100ChraftJump to review