Cycling through all six. Tap any point to stop.

Measured on six things

Wan3 AI builds freemium AI video (≤30 seconds) from text, pictures, reference footage, or documents inside the browser.

Security

HTTPS enforced and SPF present, but HSTS, CSP, frame protection, X-Content-Type-Options, Referrer-Policy, Permissions-Policy and DMARC are all missing. Checklist 3/13. No SOC 2, ISO, or bug bounty claims found.

Screenshots

1 / 1

Wan3 AI

About this app

Wan3 AI is a web AI video studio built on Wan 3.0: feed a prompt or lock the shot to owned media—reference photos, a guide take, opening and closing frames, audio, or a PDF/deck/Markdown brief—and keep characters plus packaging consistent across a clip that never exceeds thirty seconds. Extension controls reopen the result; the UI prints credit cost before render and refunds failures; Wan 3.0 sits with Veo and sibling engines on a shared pool. Trial credits may appear without a card; monthly seats start near 19.9 USD, or about 15.9 USD when billed yearly.

Key features

1Multimodal referencing: anchor shots to photos, clips, opening/closing frames, audio, or a PDF/deck/Markdown brief.
2Locked character and style consistency across multi-shot clips capped at thirty seconds.
3Shared credit pool across Wan 3.0, Veo, and other models with per-render cost shown before generation and refunds on failure.

Target customer profile

Drawn from the product itself, not from a survey.

Scout Review

Written by AI from measured evidence, scored out of 100.

Wan3 AI is a web video studio built on Wan 3.0 that lets you anchor shots to owned media — reference photos, guide takes, opening and closing frames, audio, or a document brief — and keep characters and packaging consistent across clips capped at thirty seconds. The site is fast (Lighthouse mobile 89), the docs are real, and the pricing is transparent with a no-card trial. The design is competent but undercut by small low-contrast text and a modal that fires on load. Security is the clear weak point at 3/13 header checklist points with no compliance certifications visible. Growth positioning is genuinely differentiated in a crowded category.

Measure by measure

Usability

Prompt-first hero, no-card trial credits, and a real docs site with structured navigation and a prompt guide. The value prop beyond the headline is hard to read, and the immediate modal adds friction before any content.

Design

Clean hero that puts the prompt box front and center, with a consistent violet accent. Undercut by tiny low-contrast body text, generic 3D gift-box art, and a credits modal that slams over the page on load.

Speed

Lighthouse mobile 89/100 with LCP 1.9s and CLS 0.005; desktop 85/100. TBT of 161ms shows some main-thread work but the page is fast and stable by any reasonable standard.

Security

Accessibility

Lighthouse accessibility 96/100 with one failing audit: insufficient color contrast. No WCAG statement or a11y page found. Responsive layout confirmed by successful mobile audit run.

Growth

Real problem (consistent multi-shot AI video), specific segments (influencers, product marketers), and a genuine wedge: multimodal referencing plus a shared credit pool across Wan 3.0 and Veo. Category has headroom.

Improvement areas

Security headers at 3/13 — missing HSTS, CSP, frame protection, Referrer-Policy, Permissions-Policy, and DMARC.
Body copy and feature chips too small and low-contrast; credits modal interrupts the page on load.
No accessibility statement, WCAG mention, or published security/compliance page.

Design & Usability

The design is competent and product-forward — the prompt box sits right under the headline so you see the tool, not a stock photo of someone using it. A single violet accent runs consistently from the announcement bar through the nav CTA and the generate button. What drags it down is the details: body copy and feature chips are set too small and too low-contrast to read, the gift-box illustration is generic 3D filler, and a credits modal fires over the page before you've seen anything. Usability is stronger — no-card trial credits, a real docs site with structured navigation and a prompt guide, and a clear primary path. The modal-on-load is the main friction point.

Speed & Security

Speed is a genuine strength: Lighthouse mobile 89/100, LCP 1.9s, CLS 0.005, desktop 85/100. The page loads fast and doesn't jump around. Security is the weak spot. HTTPS is enforced and SPF is set, but that's where it stops — no HSTS, no CSP, no frame protection, no X-Content-Type-Options, no Referrer-Policy, no Permissions-Policy, no DMARC, and secure cookies aren't flagged. That's 3 of 13 checklist points. No SOC 2, ISO certification, or bug bounty program appears anywhere in the collected material. For a product handling uploaded media and payment, that's thin.

Accessibility & Growth

Accessibility scores 96/100 on Lighthouse with a single failing audit for color contrast — consistent with the visual audit finding body text hard to read. No WCAG statement or dedicated accessibility page exists, and automated audits only cover part of the picture, so treat the 96 as 'obvious issues pass' rather than a full bill of health. Growth is the more interesting story: Wan3 targets a specific problem (keeping characters and packaging consistent across multi-shot clips under thirty seconds) with a concrete wedge — multimodal referencing from images, clips, frames, audio, or a PDF brief, plus a shared credit pool across Wan 3.0, Veo, and other models. The segments named (influencers, product marketers, short drama) are real and underserved. The category is crowded and moving fast, but the consistency-plus-multi-model angle is a defensible position.

Conclusion

If you're a creator or product marketer who needs consistent characters and packaging across a campaign of short clips, Wan3's multimodal referencing and shared credit pool are worth a trial — the no-card credits make that cheap to test. Go in knowing the security posture is minimal and the site's own readability needs work. The product's core promise is specific and the execution on speed and docs backs it up; the gaps are in hardening and polish, not in the idea.

Suggestions for improvement

Add HSTS, CSP, X-Content-Type-Options, Referrer-Policy, and DMARC — the header checklist is at 3/13 and these are cheap wins.
Publish a security page with data handling, storage, and any audit or pen-test status to build trust for a paid media product.
Fix the color contrast on body copy and feature chips, and delay the credits modal until after the user has seen the hero.
Add an accessibility statement and verify keyboard navigation and screen reader support beyond the automated audit.

How we score

Alternatives comparison

Named competitors, point by point. Nobody paid to appear here or to be left out.

	Wan3 AI
Multimodal reference input (images, clips, frames, audio, documents)	Yes — photos, guide takes, opening/closing frames, audio, PDF/deck/Markdown brief	Not specified in provided details
Character and style consistency across multi-shot clips	Yes — locked faces, products, and style across clips up to 30 seconds	Not specified in provided details
Multi-model access on a shared credit pool	Yes — Wan 3.0, Veo, and more share one credit pool	Not specified in provided details
Free trial without credit card	Yes — up to 50 trial credits, no card required	Not specified in provided details
Entry monthly price	$19.9/mo monthly, $15.9/mo billed annually	Not specified in provided details
Published documentation and prompt guide	Yes — structured docs with Getting Started, Usage Guides, Prompt Guide	Not specified in provided details