Cycling through all six. Tap any point to stop.

Measured on six things

Create AI video clips from text prompts or reference images with supported audio and output controls, then review before export.

Speed

Lighthouse mobile 27/100 with LCP 8.9s, TBT 20.7s and FCP 5.0s. CLS 0 is the one clean number. The site is heavy enough that mobile visitors wait nearly nine seconds for the main paint.

Screenshots

1 / 1

VEO3

About this app

VEO3 (veo3.dev) is an AI video generator that creates short video clips from text prompts or reference images. Choose a starting prompt or image, set action and camera direction, pick aspect ratio and resolution, generate with supported audio, then review the result before export. Tagline: AI Video Generator from Text and Images. Submitted by Hule.

Key features

1Text-to-video and reference-image-to-video generation with subject, motion, camera and atmosphere controls.
2Aspect ratio and resolution selection (16:9, 4s, 720p shown) with audio support where the model allows.
3Credit-based pricing across three tiers with annual billing and commercial usage rights.

Target customer profile

Drawn from the product itself, not from a survey.

Demographic

Content creators and social media marketers aged 20–35

Pain points

Need eye-catching vertical video quickly without expensive shoots

Primary needs

Fast, cheap AI-generated video clips for Reels/TikTok

Demographic

Small business owners and product marketers

Pain points

High cost of video production for product demos and ads

Primary needs

Affordable motion studies and social hooks from product images

Demographic

Filmmakers, directors, and storyboard artists

Pain points

Difficulty visualizing scenes before committing to production

Primary needs

Rapid moving storyboards to test framing, pace, and tone

Demographic

Advertising and creative agencies

Pain points

Slow client approval cycles and expensive concept testing

Primary needs

Quick visual drafts for pitches and A/B concept tests

Scout Review

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

VEO3 is a focused Veo 3 workspace with a genuinely well-crafted marketing site and a clear directional workflow — camera, action, audio, reference image — that distinguishes it from single-prompt competitors. The pricing tiers are visible and the 3-step onboarding is legible. But the fundamentals are shaky: a 27/100 mobile Lighthouse score with an 8.9-second LCP will lose most visitors before they see the product, security headers are largely absent, and the pricing FAQ answers nothing. It's a promising wedge sitting on an under-engineered foundation.

Measure by measure

Usability

Clear 3-step flow (pick source, define moment, generate and decide) and visible pricing tiers. But pricing FAQ answers render as empty dashes, and the changelog stops at a single v1.0.0 baseline.

Design

Dark editorial layout with a convincing workspace mock (16:9, 4s, 720p, subject/motion/camera controls). Dense but deliberate typography; feels more like a studio tool than a generic AI wrapper.

Speed

Lighthouse mobile 27/100 with LCP 8.9s, TBT 20.7s and FCP 5.0s. CLS 0 is the one clean number. The site is heavy enough that mobile visitors wait nearly nine seconds for the main paint.

Security

HTTPS enforced with SPF and DMARC, but no HSTS, CSP, frame protection, Referrer-Policy or Permissions-Policy — 4/12 checklist points. No SOC 2, ISO or bug bounty claims published.

Accessibility

Lighthouse accessibility 91/100 with four locales (EN/FR/DE/JA). Two real failures: unnamed buttons and color-only link distinction. No published a11y statement found.

Growth

Real problem (cheap vertical video and storyboards) with a specific wedge: camera/action/audio direction plus reference-image anchoring in one workspace. Crowded category, but the workflow focus is a defensible angle.

Improvement areas

Mobile performance: 27/100 Lighthouse with 8.9s LCP and 20.7s TBT.
Security headers: no HSTS, CSP, frame protection, Referrer-Policy or Permissions-Policy.
Pricing FAQ renders six questions with no answers, leaving credit and billing mechanics unclear.

Design & Usability

The site reads like a design studio's internal wiki: dark, typographic, and structured around a mock workspace panel showing aspect ratio, duration, resolution and subject/motion/camera controls. That's a stronger first impression than most AI-video landing pages, which tend to be a hero video and three gradient cards. The 3-step flow (pick a starting point, define the moment, generate and decide) is genuinely clear, and the pricing page surfaces three tiers without hiding behind 'contact sales'. The cracks show in the details: the pricing FAQ lists six questions and answers none of them, and the changelog contains exactly one entry, a v1.0.0 baseline from 2025-05-20. For a product asking $9.9–$59.9/month, unanswered billing questions about credit expiry and pay-as-you-go are a real friction point, not a cosmetic one.

Speed & Security

This is where the product is weakest. Lighthouse mobile scores 27/100 with an 8.9-second LCP, 5.0-second FCP and a brutal 20.7-second total blocking time — the page is heavy enough that mobile visitors will bounce before the hero renders. CLS of 0 is the sole bright spot. On security, HTTPS is enforced and SPF/DMARC are configured, but the header probe found no HSTS, no CSP, no frame protection, no X-Content-Type-Options, no Referrer-Policy and no Permissions-Policy — 4 of 12 checklist points. There's no published SOC 2, ISO 27001, pen-test summary or bug bounty. For a tool that accepts uploaded reference images and generates commercial-use video, the absence of a security page and hardened headers is a gap worth closing before agencies start sending client assets through it.

Accessibility & Growth

Accessibility lands at 91/100 on Lighthouse with four supported locales (English, French, German, Japanese), which is above average for this category. Two concrete failures remain: buttons without accessible names and links distinguished by color alone — both cheap to fix and both directly relevant to a product whose pitch is visual judgment. Growth is the more interesting question. The problem is real: creators and agencies need cheap vertical video and moving storyboards without a shoot. The wedge is specific — a workspace that lets you direct camera movement, subject action and audio alongside a reference image, rather than a single prompt box. Against Runway, Pika and Sora, that workflow focus is the differentiator, but the category is crowded and the moat is thin: the underlying model is Google's, and the domain is only months old. Execution and pricing clarity will decide whether this becomes a tool or a tab someone closed.

Conclusion

If you're evaluating VEO3, judge it on the workflow, not the landing page: the camera/action/audio direction and reference-image anchoring are the actual product, and they're worth testing against Runway and Pika on the same prompt. Before committing to an annual plan, ask support the questions the FAQ doesn't answer — credit expiry, pay-as-you-go mechanics, and what happens to unused credits on cancellation. And if you're the maker: fix the LCP, add the missing security headers, and fill in those six FAQ answers. None of those are hard, and all three are currently costing you conversions.

Suggestions for improvement

Cut mobile LCP from 8.9s — compress hero assets, defer non-critical JS, and target sub-2.5s.
Add HSTS, CSP, X-Content-Type-Options, Referrer-Policy and Permissions-Policy headers.
Fill in the six pricing FAQ answers, especially credit expiry and pay-as-you-go mechanics.
Give every button an accessible name and stop relying on color alone to distinguish links.

How we score

Alternatives comparison

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

	VEO3	Runway ML	Pika Labs	Sora by OpenAI
Text-to-video generation	Yes — powered by Google Veo 3	Yes — multiple Gen models	Yes — text and image inputs	Yes — limited access
Reference image to video	Yes — anchors subject and composition	Yes — image-to-video supported	Yes — image-to-video supported	Limited
Camera and action direction controls	Yes — subject, motion, camera, atmosphere	Yes — advanced editing and camera controls	Partial — prompt-driven	Prompt-driven
Audio generation	Yes — supported Veo 3 audio	Partial	Limited	Limited
Pricing model	Credits: $9.9–$59.9/mo annual	Tiered subscription credits	Freemium + subscription	Included in ChatGPT plans
Collaboration and team tools	Not advertised	Yes — collaboration tools	Limited	Limited