Dreamflow Review: AI App Builder, Pricing & My Test Results

Dreamflow is an AI app builder that makes Flutter apps. You describe an app in a prompt box, an agent plans and writes it, and you see a live preview next to the code.

The plans are Free with 10 credits, Hobby at $20 a month with 100 credits and Pro at $90 with 500. I used Free, because the aim of this test was to go as far as the free plan allows and report what it blocks.

I tested the Free plan, not a paid one, and Free is limited: 10 credits for life, web publishing only and a locked code download. Every number here comes from my own tests and my own monitors, and every app builder gets the same brief, checks and attacks. Six tests are graded. Four are not: payments and emails need your own keys, the twelve changes needed more credits than Free had left, the code download is locked and support has not answered.

Where a test could not run on the Free plan, the section says what stopped it and does not guess.

Dreamflow has a Free plan, then $20 Hobby and $90 Pro, each with more credits. Prices read on 1 November 2026. The pricing page shows one monthly price for each plan.

The pricing page lists no top-ups, so when the credits run out the page offers no way to buy more. What a finished app costs is not something I can tell you.

Dreamflow pricing page with Free at $0 and 10 credits, Hobby at $20 a month with 100 credits

Here is the number you will ask about first. In total I paid $0. The Free plan gives 10 credits for life, with no card needed. The credits are what ran out.

Dreamflow Plan and Billing page showing a banner that the projected use is 262 credits, the Free plan

Dreamflow's own banner then said I was on course to use 262 credits this period. Hobby has 100 and Pro has 500, so I cannot tell you what a finished version of this app would cost.

The build step was the costly one: Continue building alone used about 60% of the Free allowance. Even my one-letter message, cut short by a typing fault of mine, cost a small part of a credit. If you build on Free, expect to run out after about four prompts.

The first prompt gave a plan, not an app. About six minutes after I pressed send, the agent showed an architecture plan with a feature list and twelve steps, and the preview read Project has 1 error. A second prompt, Continue building, took about eight minutes and ended with a running app and no errors.

Dreamflow editor with the Fitting Room brief in the agent chat, an Architecture Plan listing key features

I published it on a share.dreamflow.app address about 22 minutes after the brief, then spent one more prompt on the classes it had invented, and published again at about 35 minutes.

The target was a live app with 20 of the 24 checks passing, and the app ended on 10, so you would still have work to do. In the middle, my own typing fault sent the agent a one-letter message, and it asked what I wanted next.

A 1, because the checks never got near the passing line.

Not mostly. 10 of 24 checks passed, 11 failed and 3 were not run. Six more, the payments and the two emails, need keys that are mine to supply, so I left them out. A member can sign up, book, see the booking and cancel it.

The first live version invented three classes, a Jordan Instructor and two studios. My fix prompt put the three real classes in. The admin screen is a stub, with no way to add, edit or delete a class, no booking list, no CSV and no Ask the studio form anywhere. The studio address, phone number and prices are on no screen I opened.

When a member cancelled, the places went back up, from 5 to 6. The app also added levels (Beginner, All levels, Advanced) and a studio label that the brief never asked for. The capacity rule needs six member accounts, which I did not create, so that check, member against member and marking attendance were not run.

Dreamflow test app member timetable listing Reformer Basics with Lena Park, 8 spots, Mat Flow with Tomas Reyes

Dreamflow test app admin screen with the text Welcome Dana Okafor, a note that admin tooling is next

I could not run the twelve changes, so this test is not graded. The Free plan gives 10 credits for life, and the build and one fix used 87% of them.

The one change I did run, replacing the invented classes, cost under two credits and worked first time.

At the rate of my one fix, the other eleven changes would have needed about 15 credits, and Free had under 2 left. If you plan many edits, you will need a paid plan.

Dreamflow editor with a phone preview of the Fitting Room app on the left and the agent chat

Publishing is one click on the Web tab and took under four minutes. The app answered on a free share.dreamflow.app address, and Update Deployment pushes a change the same way. The panel offered no custom domain.

Dreamflow Publish panel on the Android tab saying Available only on paid plans

The speed score looks good, but the page shows a loading screen first and the app after. I did not count layout faults or try a real phone.

Every day I also run the Dreamflow speed test and the Dreamflow security checks.

A review tells you what happened on one day. This tells you what happens every day after. My own monitor opens the app every five minutes for 30 days, from 27 October to 26 November 2026, and records whether it answered and how fast.

Dreamflow hosts the app on its Free plan. Whether a free address stays up for 30 days is something the record will show.

Every check so far is on the Dreamflow uptime record.

Sign-up and log-in work, but the roles are only labels. The sign-up form lets anyone pick Member, Instructor or Admin, and I made an admin account with a new email in one step. The first screen also prints the demo password of every role.

Dreamflow test app Create account form saying Local demo only, no backend yet, with a Role list open showing Member

There is no server. The agent saved the users, the bookings and the session in the browser, and the password is stored in a form anyone can read back. So my two attacks had nothing to attack.

Dreamflow test app first screen with the title Fitting Room, three buttons for member, instructor and admin

It survives a reload, in one browser. One of my six data checks passed. The admin screen shows no data, so viewing, filtering, editing, deleting and exporting all failed.

Dreamflow has Firebase and Supabase panels for a real database. The agent did not use either, and I did not connect one, because that needs an account that is yours.

Browser storage never leaves the browser. I did not open a second device, but a booking made on one phone would not be on another. I also did not run the republish and colour-change checks, because only about a tenth of the credits were left.

Dreamflow test app My bookings screen listing Mat Strength with Jordan Instructor, East Austin Studio

I did not run payments or emails, because Stripe and an email service need keys that are yours to supply. The six checks that depend on them are left out, and this test is not graded. The password reset screen says a link was sent, but no email exists.

Dreamflow test app sign-in screen with a dark bar at the bottom reading Password reset link (demo) sent.

The brief asks for a $95 five-class pass. The app has no screen where a member can buy a pass, and booking a class never mentioned one. If you sell passes, you would have to add that yourself.

Exporting means downloading your app so it can run somewhere else. On the Free plan Download Code is locked: the menu item is greyed out and says Need plan upgrade. The pricing page lists code export on Hobby and Pro.

Dreamflow project menu with Download Code greyed out and the words Need plan upgrade.

I did not read the terms, so I cannot say who owns the app you build. Check this before you pay if your app must leave one day.

I emailed one question to the support address that serves Dreamflow, and an automatic answer came 27 seconds later. No person had answered when I wrote this review, so support is not graded. I will update the review when one does.

The Free plan has nothing to cancel. The account page has a Delete Account button, which I did not press. I never paid, so I could not test a paid cancellation.

Dreamflow has a Trustpilot TrustScore of 2.3 from 8 reviews, read on 28 October 2026, and all of them give one star. That is a very small number, and one of the eight describes FlutterFlow, not Dreamflow. G2 would not load for me.

The complaints are about cancelling and being charged after, credits used up fixing the platform's own errors, and slow support and replies written by bots. One reviewer wrote "nothing works anymore once you start using Supabase". I did not connect Supabase, so I cannot confirm that. I never paid, so I cannot test the billing complaints.

If this were my studio, I would not run it on the app Dreamflow's agent built for me. It looks right and works for one person in one browser. The admin is empty, anyone can become the admin, and no bookings are shared between devices.

It suits people who want a fast, free Flutter demo to show an idea, and who are ready to add a real backend and pay for the export. It suits you less if you want the agent to deliver a working booking system. This was the Free plan, so I could not try payments, run the twelve changes on 10 credits, download the code or test a paid cancellation, and support has not answered.

The six graded tests average 2.0 out of 5, and I rate Dreamflow 2.0. Publishing and pricing are the highest grades, at 3, and two tests score 1.

See the Dreamflow test results and how I rate.