Abstract
Vidu has a real backlash around removed daily check-ins, broken ad rewards, expiring credits, strict moderation, and wasted credits on bad generations, but the incumbent still has 1,868,923 installs, a 4.8★ lifetime rating, and users praising output quality; a solo builder can attack the credit-policy gap, not realistically outbuild the core video model.
Background
Vidu - AI Video Generator turns text and images into short AI videos with audio, remixing, references, and community-style creation; the Play listing is free, with 1,868,923 installs as of 2026-08-03 and a 4.8★ current lifetime average.
What it does
Prompt-to-video and image-to-video creation for social clips, characters, dance videos, story scenes, and remixable community ideas.
What changed
Recent reviewers repeatedly say the daily login/check-in credit reward disappeared, while ad-based earning is unreliable or too limited.
Why it is soft
The anger is about fairness, credit loss, and support silence, not about people losing interest in AI video creation itself.
Strengths
This is not a weak product with easy users to steal: many reviewers still call Vidu easy, high quality, realistic, fun, and better than other AI video generators, and several positive reviews specifically like that credits can still be earned through tasks or ads.
Output quality
Even frustrated users often separate the credit problem from the fact that the videos can look impressive when the generation works.
Proven demand
The install base shows broad consumer curiosity, and the reviews show people are willing to watch ads, do tasks, or pay when the value feels fair.
Habit loop
Daily credit collection became part of the product habit, which is exactly why removing it created such visible disappointment.
Market gap
The gap is not “make AI video” in general; it is “make AI video with predictable credits, transparent refunds, and fewer surprise policy changes.” The complaint clusters below are ordered by how prominent they appear in the provided reviews.
Daily credits removed
The most common complaint is the missing daily login or check-in bonus, which users describe as the reason they kept opening the app.
Ad rewards and credit loss
A visible second cluster says ads do not always pay out, credits expire or vanish, and users feel charged for unusable generations.
Generation failures and limits
A smaller but important cluster mentions queues, crashes, short outputs, sloppy generations, and frustration when obvious errors still cost credits.
Moderation and support
Some users object to strict restrictions, teen limits, asset deletion, and support emails that allegedly get no reply.
The daily check-in task for earning free points has been discontinued. AI apps are taking things away, one by one. Nothing lasts forever. But we can still earn free points for generating a video, liking 5 posts, and publishing a video. Now the "watch ads for points" is back again.
I had to change my review. Not only do credits expire quickly, but the app keeps removing or changing ways to earn credits. The login reward was removed completely. Now it added watching videos as a way to earn credits in addition to other rewards. Again, credits will vanish daily, so spend them or lose them. The generations have gotten very sloppy, so you will undoubtedly waste credits trying to get something that doesn't look like nightmare fuel.
was 5 stars however, they have removed the daily check in which allowed you to generate a couple of videos a day. only way to earn is to watch ads which doesn't work! the generated videos are only 8 seconds check other apps that create longer videos for the same monthly fee! email support and don't get a reply.
- Make credits non-expiring and auditable. Show every earn, spend, refund, and failed job in a ledger so users never wonder where points went.
- Refund failed or unusable generations automatically. The reviews tolerate imperfect AI, but not paying credits for crashes, obvious errors, or moderation blocks.
- Keep the daily habit, but cap it honestly. A small daily grant plus optional ads is easier to trust than silently removing the feature that trained users to return.
Build complexity
Estimated, not derived: one developer, part-time, no funding, and no paid acquisition. This assumes a wrapper around existing AI video generation services, not training a model; matching Vidu’s best output quality from scratch is not realistic for a solo builder.
| Workstream | Weeks |
|---|---|
| Mobile app shell, onboarding, gallery | 2.0 |
| Prompt/image upload, job queue, status screens | 2.0 |
| Video generation integration, retries, downloads | 2.5 |
| Credit ledger, subscriptions, ad-reward fallback | 1.5 |
| Privacy controls, basic remix, reporting | 1.5 |
| Moderation, QA, Play listing, support flows | 2.5 |
| Total | 12.0 |
Expected revenue
Estimated, not derived: with no paid acquisition, assume a simple $9/month subscription plus a transparent monthly credit allowance; the month-12 funnel below is gross MRR before generation/API costs, refunds, store fees, and taxes.
| Month 12, monthly | Users | Assumption |
|---|---|---|
| Installs | 3,000 | Organic only |
| Activated | 900 | 30% |
| Retained at 30d | 450 | 15% |
| Paid subscribers | 167 | $9/month |
| Revenue | $1,500 | Gross MRR |
Biggest challenges
The hard part is not building a prompt box; it is earning store trust against a highly rated incumbent, paying for expensive generations before revenue is stable, and writing moderation/support policies that do not create the same resentment.
Cold-start trust
Vidu has a large install base and a 4.8★ lifetime average; a new app starts with no proof that its credits, refunds, or outputs are better.
Inference cost risk
Free daily credits are the demanded feature, but every generation has a real cost, so generosity can quickly become an unfunded liability.
Moderation burden
Reviews complain both about restrictions and about asset deletion; a solo developer still has to satisfy users, model providers, and Play policy.
AI angle
There is a real AI angle, but it is enabling rather than differentiating: off-the-shelf AI video generation can make a solo-built wrapper possible, while AI classifiers can help flag failed generations for automatic credit refunds; it does not give a solo builder a durable advantage over Vidu’s core model quality.
Useful AI leverage
Use AI to score motion glitches, obvious prompt misses, and moderation-blocked outputs so the app can refund credits without manual support.
Not a moat
The incumbent already is an AI video product; the solo moat is transparent economics and trust, not a better foundation model.
Conclusion
Skip it. Build only a narrow fair-credit experiment if you already know app distribution; the reviews expose a real trust gap, but a part-time solo builder is fighting a strong incumbent, high generation costs, and a crowded AI-video category for modest likely MRR.
12 weeks
Time to v1
$9/mo
Subscription + credits
$1,500
Revenue at month 12
4.8★
Incumbent still trusted