Cycling through all six. Tap any point to stop.

Measured on six things

Independent AI image and video creation workspace.

Security

HTTPS and HSTS are in place, but that's where it stops: no CSP, no frame protection, no Referrer-Policy, no SPF, no DMARC. 4/13 on the header checklist with no published compliance or audit claims.

Screenshots

1 / 1

Flux 3 Video Generator

About this app

Flux3 is an independent AI image and video creation workspace offering text-to-image, reference remix, text-to-video, and image-to-video workflows, with controls for aspect ratio, style, resolution, and duration. Account required. Not affiliated with or endorsed by Black Forest Labs.

Key features

1Text-to-image and reference remix workflows with aspect ratio, style, and resolution controls
2Text-to-video and image-to-video generation with duration and delivery format settings
3Credit-based access with cost shown before submit and 40 launch credits on signup

Target customer profile

Drawn from the product itself, not from a survey.

Scout Review

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

Flux 3 Video Generator is a well-written, well-typeset front door to a product that hasn't shown you its rooms. The copy is unusually thoughtful for the category — it talks about lens choice and material treatment instead of 'unleash your creativity' — and the three-step workflow with pre-submit credit costs is a sensible design. Speed is genuinely good. But the homepage contains zero actual generated output, there's no documentation or demo, the security headers are half-configured with no email authentication, and the accessibility failures hit the interactive controls directly. It's an early-stage wrapper around FLUX models with a clear workflow idea and no moat. Worth a look if you need still-to-motion continuity; not yet worth committing a production budget to.

Measure by measure

Usability

Clear three-step loop (choose start, direct scene, generate/refine), credit cost shown before submit, 40 launch credits to try it. But no docs, no FAQ, no public demo — you must sign up to see anything work.

Design

Clean, editorial landing page with disciplined typography and a convincing console mockup — but the 'OUTPUT' panel is an empty placeholder and the page shows zero real generated images or clips. Marketing craft exceeds product proof.

Speed

Lighthouse 87 mobile / 97 desktop, CLS 0 and TBT 46ms are genuinely tight. The one blemish is a 4.0s mobile LCP, well past the 2.5s 'good' line — likely a heavy hero asset.

Security

HTTPS and HSTS are in place, but that's where it stops: no CSP, no frame protection, no Referrer-Policy, no SPF, no DMARC. 4/13 on the header checklist with no published compliance or audit claims.

Accessibility

Lighthouse 87/100, but the failures are the ones that actually block people: buttons with no accessible name, invalid ARIA values, and contrast that doesn't pass. No WCAG statement, English-only.

Growth

Real, growing demand for text/image-to-video, and a specific wedge: keep the brief intact across still and motion. But it's a thin wrapper over FLUX models with no moat and a crowded field.

Improvement areas

No real generated output shown anywhere on the homepage — placeholder panel only
Missing security headers (CSP, frame protection, Referrer-Policy) and no SPF/DMARC
Accessibility failures on core controls: unnamed buttons, invalid ARIA, low contrast
No documentation, FAQ, or public demo — account required before any evaluation

Design & Usability

The landing page is genuinely well-made: consistent type scale, numbered sections, a console mockup that shows exactly what the product does. The problem is that it shows what the product does without showing what it produces — the output panel is a placeholder, and there isn't a single real generated frame or clip anywhere on the homepage. For a visual generation tool, that's the one piece of proof that matters. Usability is a similar story: the three-step loop is clearly explained, credit cost is surfaced before you submit, and 40 launch credits lower the barrier. But there's no documentation, no FAQ, no public demo, and no way to evaluate output quality without creating an account. You're asked to trust the copy.

Speed & Security

Speed is the strongest measured dimension. Lighthouse gives 87 on mobile and 97 on desktop, with CLS at 0 and TBT at 46ms — the page is stable and responsive under interaction. The single weak point is a 4.0s mobile LCP, which suggests a heavy hero image or font blocking first paint; fixing that would push this into the high 80s. Security is the opposite story. HTTPS and HSTS are correctly configured, but the header probe found no CSP, no frame protection, no X-Content-Type-Options, no Referrer-Policy, and no Permissions-Policy — 4 of 13 checklist points. DNS shows no SPF and no DMARC, which leaves the domain open to email spoofing. There's no security page, no compliance claim, and no bug bounty. For a tool that stores user prompts and generated media, that's thin.

Accessibility & Growth

Accessibility lands at 87/100 on automated checks, but the three failures are the ones that matter most in practice: buttons without accessible names, invalid ARIA attribute values, and insufficient color contrast. Those are cheap to fix and currently block screen reader users from operating the core generator. There's no accessibility statement and no language support beyond English. Growth is a judgement on the idea: text-to-video and image-to-video are genuinely expanding categories, and the 'carry the brief from still to motion' positioning is a real, specific wedge rather than a generic 'AI video' claim. But the product is a workspace layered on top of third-party FLUX models, with no proprietary model, no network effects, and no data advantage. The domain is roughly three months old, so there's no traction to point at either way. It can win a niche on workflow polish; it cannot build a moat on it.

Conclusion

The honest read: this is a competent, early product with a real workflow thesis and a marketing site that oversells what it can currently prove. The fixes are concrete and mostly cheap — publish real outputs, write a docs page, add the missing security headers and DMARC, name your buttons. Do those four things and the score moves meaningfully. Until then, treat the 40 launch credits as the entire evaluation budget and decide from what you actually generate, not from what the landing page says you will.

Suggestions for improvement

Put real generated images and video clips on the homepage — the placeholder output panel is the single biggest credibility gap.
Add CSP, X-Content-Type-Options, Referrer-Policy, Permissions-Policy, plus SPF and DMARC records to the domain.
Fix the three Lighthouse a11y failures: accessible names on buttons, valid ARIA values, and passing contrast ratios.
Publish a docs/FAQ page and a public sample gallery so users can judge output quality before creating an account.

How we score

Alternatives comparison

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

	Flux 3 Video Generator
Text-to-image generation	Yes — with aspect ratio, style, resolution controls	Not specified	Not specified	Not specified
Text-to-video generation	Yes — duration and format controls	Not specified	Not specified	Not specified
Image-to-video / reference remix	Yes — reference remix and image-to-video	Not specified	Not specified	Not specified
Free tier / trial credits	40 launch credits on signup	Not specified	Not specified	Not specified
Public output gallery / demo	No — placeholder panel, account required	Not specified	Not specified	Not specified
Documentation / help center	None found on site	Not specified	Not specified	Not specified