Rocket Avaliações 93

Clearly, this is a money trap. After 1 hours of simple building the architecture of an app, the first layer of subscription is not enough.

After a little bit of architectural work, the AI burns its "Token" and you can't progress anymore.

Off course, there is no free trial and can't get a refund after 1 hours of work.

Scam!

I have been building a small app which takes 15M tokens, each time you do or add something it breaks down another setting. Fixing it costs you tokens!!!
The idea is really great and maybe, just maybe after the beta it will be better but at this point it just eats your money away!!! Just frustrating!! as still all functions aren't working even after requesting to fix this several times. 15M tokens for nothing!!!!!!!!!!!!

The app itself and the AI. 5/5.
It allows you to create apps previews of what you want in a matter of minutes, the AI works really well.

The implementation in a real app. 3/5.
Here is where problems start. The final app wouldn't work as the preview cause it'll probably need the use of Supabase and things that aren't obvious for the non-programmers. So what looked like an easy task turned to be much difficult and more frustrating than expected in order to make it work as a downloadable app.

The subscription plans 2/5.
I think 25 dollars a month for a subscription with limited tokens is too expensive. The tokens will run out very fast the moment you try to correct things that don't work or if you want to do various small changes. I think it should give much higher number of tokens with the 25 dollar plan. Most of my frustration came from running out of tokens and having to pay another month just to fixed problems that didn't appear on the preview of the app.

The Help Service 5/5
Here the app really shines, it has a very good human service that helped me a lot to fix the problems that I had until were solved, certainly the best help service I ever had. So I guess there is a great team behind. Because of that, I think this app has a lot of potential on becoming a reference in AI app making. I wish them the best.

Overall, making the app work wasn't as easy as it seemed but I learned quite a lot about the world of AI and app making and now I'm more interested in learning more about it in the future

Rocket is working on an extremely complicated website with several API’s, databases and payment interfaces. Support has been amazing in helping me overcome major obstacles that seemed insurmountable. The finish line is just a short distance away.

I am not a developer but I must say this Rocket is the best AI app developer in the world the support it outstanding from start to the end they ar always responding to my queries and very supportive in my apps development stages. Rock on Rocket

I’m writing this review to document a full, real usage session with Rocket, not a first impression like the other day.

Timeline:
From Jan 6, 2026 – 1:23 PM to 9:01 PM (almost 8 hours).
I personally spent around 80% of that time actively working with the system.

What I tried to build

A professional informational website (not a SaaS, not a marketplace, not a complex platform):

Single-page website

Clean, minimalist design

Fully functional navigation

A real admin panel able to control 100% of the frontend

Editable text, colours, sizes, images, videos

GDPR-compliant contact form

Production-ready, not a demo

I provided an extremely detailed, explicit prompt, with clear acceptance criteria and no ambiguity.

Token usage & cost (important)

I started the day with ~3.4M tokens (including 1M added earlier by support as a goodwill gesture).

During the session, those tokens were completely exhausted.

At 5:13 PM, despite mounting issues, I bought $20 worth of additional tokens (~5M) because I genuinely wanted to give the system a fair chance.

At the end of the session, about 3.8M tokens remained, but I was mentally exhausted from constant firefighting.

A large portion of the tokens were consumed trying to:

restate requirements already clearly given

fix partial or broken changes

deal with regressions

address errors introduced by previous “fixes”

What actually happened

To be fair: some progress was made.
The website did improve compared to earlier attempts.

However, the experience was extremely unstable and inefficient:

Features repeatedly marked as “Completed” were not actually completed

Many changes were partial, inconsistent, or broken

Fixing one issue often introduced new errors

The admin panel required dozens of iterations to approach basic usability

Live Preview initially did not work, then worked partially, then broke again

Sections were hardcoded when they were supposed to be editable

Admin and frontend section names didn’t match

Media uploads broke the build

Replacing a simple logo triggered massive code corruption

Eventually the platform hit Next.js 15 / SWC build system errors that Rocket itself could not fix inside the platform

At one point, Rocket told me to run terminal commands — then admitted I cannot run them on Rocket and that it was a platform infrastructure issue requiring support intervention.

Support & refund

I exchanged two emails with support during this process.

Rocket eventually refunded the $25 subscription I paid two days earlier.

They also added 1M extra tokens and left the remaining tokens in my account.

I appreciate the refund and the polite tone from support.

That said, time is more valuable than tokens — and the time cost here was very high.

The core problem

Rocket is marketed as a system capable of building complex apps quickly.

My experience shows that even for a non-complex professional website, the platform:

requires constant manual correction

burns through tokens very fast

struggles with consistency and system-wide thinking

does not reliably validate whether a task is actually done before claiming completion

After nearly 8 hours, millions of tokens, and significant mental effort, the result was still not production-ready.

Final thoughts

I truly wanted Rocket to work. I acted in good faith, invested time, patience, and additional money.

But in its current state, for serious professional work, Rocket feels unreliable, token-inefficient, and extremely time-consuming.

A great review should be earned by a great job.
Unfortunately, this experience did not reach that level.

I hope Rocket improves — but users should go in with very realistic expectations.

The Rocket team appears to be very responsive to an issue(s) that you may be experiencing and really strive to bring forth a resolution(s) in assisting with the application development process.

Tuve una muy buena experiencia con el equipo de soporte de Rocket.new. Enfrentamos algunos problemas técnicos complejos en nuestro segundo proyecto y, como siempre, el equipo de soporte estuvo disponible y comprometido en ayudarnos a resolverlos. Se nota el profesionalismo y la preocupacion genuina porque el proyecto sea un éxito. Muy agradecido del apoyo, especialmente de Harsh, quien nos acompañó durante todo el proceso con paciencia, buen criterio tecnico y siempre atento a nuestros avances.

Rocket is genuinely impressive. I used it to build a complex, multi-feature application with real data models, authentication flows, and business logic—and it delivered far beyond what I expected from a prompt-driven platform. The speed is obvious, but what stood out was how structured the output was: clean architecture, sensible defaults, and a strong grasp of real-world product requirements. This isn’t a toy or a demo generator—it’s a serious acceleration tool for turning ideas into working software fast. Rocket.new materially compressed weeks of work into hours and cut our development budget by 80%.

Just wanted to say how much I appreciate the support I have received from the team. My app was more complicated than the norm, and they helped me every step of the way to get it sorted and launched. Very impressed with their commitment to see their users work through to completion. I highly recommend Rocket for this reason.

The program is powerful, but i quite literally used 10million tokens to fix a simple pdf reading problem in my app, only from going from partial reads to none, and still no fix... and considering im progressing backwards i have no confidence that spending another 10 million tokens will fix that.

might redo my review at a later time, but at this time the frustration with the AI saying it fixed it without actually fixing, implementing code, and not actually implementing it, and having to spend tokens on errors the ai makes can make the creation process much more expensive than you first think it will be.

Was marketed to upper level and different, the 1 million tokens is a complete sham. It basically just built the web app and after that I could send 3 chats before the notification about zero tokens started appearing. Just imagine purchasing 25 million tokens for 50 dollars that means at best you will get 15 maybe 17 chats out of 50 dollars.

After a million attempts and errors. My app was almost ready to go then another prompt destroyed the entire app and filled it with errors. I complained and tokens were refunded a few times.

Still didn’t fix anything so customer service asked me for my account info to check it.

Rolled it back to an old old version with stuff I didn’t want to retrieve. Now I can’t log into my account and recovery email isn’t being sent.

Customer service responded with “ here, use my account “ I said no.. can you help me recover my account.

2 weeks later still can’t log into my account and the refunded tokens disappeared.

Not worth your time or money.

Customer service from Raaj was very professional, thank you.

Greatest investment I've made so far. This app is turning my dreams and goals to reality. Amazing customer service, loyal developers, and great people to work with. Very supportive in all aspects of fixing and solving issues to keep you focused on your goals. Very thankful to have stumbled across this game changer. Thanks for the Amazing platform that gives people an opportunity to create their own success.

Tested both this and base 44 simultaneously.

Rocket is better at building UI, from your initial prompt will build great UI and general functionality this kind of tricks you into buying a full subscription with them. After your initial prompt any minor tweaks through prompts the AI falls apart and starts draining tokens. I was on the top tier and it drained about 7m tokens trying to add staff data points to the drop down menu. Great to build an initial template and then build manually off, but if you wanted that you would have just used Bubble.io

The product is still a beta and they've raised significant seed funding so in 6 months I can imagine the product will greatly surpass base 44 however in its current state it's not a product I would continue my subscription with. I imagine it's been released in its current state as a way to bring in revenue so they don't burn money. However it may have been a bad choice as I can see this product being incredible in a couple of months but it will take a lot of positive reviews for me to pick this back up.

pretty awful , dont recommend it

Awesome builder - modern, fast and insanely easy to use.
Rocket generates a lot from a single prompt, and after testing many other builders, this is the one I’m sticking with.

Their Discord community is also great - friendly, active, and genuinely helpful whether you need support or just want to chat.

Highly recommend. Rocket is absolutely worth trying.

You waste a few credits getting the swing of things and you do need a lot more than 5 million credits to make a full app. I would say 50 million would be perfect for an incredibly high end functional app.

Once you've got the hang of the system and the way you prompt engineer you'll have yourself a fully functional front end and back end within a week. Its so much quicker than writing an app by hand in flutter. I would recommend learning the programming language so that you can be good with the prompt engineering however

I haven't needed to program anything myself yet. It does it all for you and efficiently. So much cheaper than paying to have such a high end app made. 50 million credits is 200$ which is nothing considering the level to which the app you get is built to.

Excellent application for developers who don't know much about programming, 10/10 experience. Support also very attentive and quick responses.