The first thing I wanted to know about Rork was not whether it could generate an app. It was whether the thing it generated would be something I could fix myself later. A preview that runs in a browser is a demo. A project I can open in an editor is a tool.
Downloading a Rork build settles that question quickly: what lands on your machine is an ordinary Expo project. There is an app/ directory, a package.json, and npx expo start just works. Once you know that, the scope and the limits of this platform become much easier to reason about.
Here is what Rork actually produces, which decision you need to make before you start building, and how far you can get without paying.
You get an Expo project, not a locked-in app
Rork takes a written description and generates a React Native + Expo project. You preview it in the browser, add follow-up prompts to change things, and iterate until it matches what you had in mind.
What matters is that the output is not trapped inside the platform. Download it and you get this:
my-app/
├── app/ # expo-router file-based routing
│ ├── _layout.tsx
│ ├── index.tsx
│ └── (tabs)/
├── components/
├── assets/
├── app.json # Expo configuration
├── package.json
└── tsconfig.json
Run npm install, then npx expo start, and you are doing ordinary Expo development. Rork is less an app-making machine than a colleague who writes the entire first implementation for you.
That distinction changes how you evaluate the tool. With builders that emit a proprietary format, hitting a wall means you are stuck. With an Expo project, hitting a wall means you open an editor and keep going.
Prompts work best at roughly this level of detail:
Build an app that tracks daily water intake.
- Home screen shows today's total (ml) and a progress ring toward a goal
- A + button adds 100 / 200 / 500ml; long-press opens manual entry
- A history tab shows a bar chart of the last 30 days
- Store data on-device (no server)
The vaguer the prompt, the more generic the output. Writing the requirements does not go away. What goes away is the hours between having requirements and having something that runs.
React Native or Rork Max — decide on Android and web first
Rork comes in two flavors with genuinely different characters. Choosing late is expensive.
| Aspect | Rork (React Native) | Rork Max |
|---|---|---|
| Generated code | React Native + Expo (TypeScript) | Native Swift / SwiftUI |
| Ships to | iOS, Android, web | Apple platforms only |
| Target devices | iPhone, Android phones, browsers | iPhone, iPad, Apple Watch, Apple TV, Vision Pro, iMessage |
| Build environment | Expo build infrastructure | Compiled on a cloud Mac (no Mac required locally) |
| Platform reach | Common app features and native modules | ARKit / LiDAR, Metal, widgets, Dynamic Island, Live Activities, HealthKit, NFC, App Clips and more |
| Pricing | Free tier plus credit-based usage | Around $200 per month |
You do not need to compare feature lists. One question decides it: are you shipping to Android or the web? If yes, take the React Native version. If you are committing to Apple platforms, Rork Max gives you reach that cross-platform code cannot match. Rork Max cannot output Android or web builds at all.
Rork Max launched in February 2026 and was reported to reach $1.5M in annual recurring revenue within three days. Its generation layer is described as running on Claude Code and Claude Opus 4.6. For where native generation still leaves manual work behind, see where Rork Max native generation falls short.
Two hundred dollars a month is not a small commitment for an indie developer. Work out how many apps, over what period, would justify it before you subscribe. A full plan breakdown lives in the Rork pricing comparison.
How far the free tier actually goes
The free tier runs at roughly five prompts per week. Five disappears instantly if you generate from scratch, dislike the result, and start over.
A more realistic allocation is one prompt for the initial build and two or three for revisions. That is exactly why it pays to specify screens and interactions properly in the first prompt.
For device testing, Rork Companion lets you install a generated app on your own iPhone without a paid Apple Developer account. Simulators and real hardware are not the same experience — scroll feel and tap targets both read differently in your hand. Being able to check that before paying anything is a meaningful advantage. The walkthrough is in testing on a real device with Rork Companion.
Where "no code" stops
Rork is marketed as letting you build apps without writing code. That is accurate as an entry point. It is not a promise that holds all the way to the App Store.
In my own use, the boundary showed up in roughly three bands:
- Usually solvable by prompting — navigation, lists, forms, local persistence, basic animation
- Where you start wanting to edit — fine layout adjustments, auth against an existing API, error-state behavior, performance tuning
- Faster to write yourself — intricate platform-specific work, anything sensitive to third-party SDK initialization order
When you reach the third band, the fact that you hold an Expo project is what saves the project. You open the file and fix it.
The safer assumption is that you will read some code eventually. You do not need to write it fluently, but you want to be in a position where you or someone you trust can open the file when the prompts stop converging. Deciding in advance where you take over saves hours of circling back with the same instruction phrased differently.
What the funding history tells you
Whether a tool will still exist next year is a fair thing to weigh. Rork's reported position:
- Founded by Levan Kvirkvelia and Daniel Dhawan
- An early $2.8M round from a16z
- A $15M seed round led by Left Lane Capital, announced on April 9, 2026
- The acquisition of app builder Paperline, disclosed in the same announcement and described as an engineering talent acquisition
- Roughly 743,000 monthly visits with a reported 85% growth rate
None of those numbers make your app ship. What they inform is a narrower question: are you investing time in something likely to be maintained? On that question, continued funding and acquisition activity are worth noting.
The reception is not uniformly positive, though. ICON POLLS scores Rork 3.8 out of 5, citing pricing, the credit consumption experience, and occasional instability. Going in with calibrated expectations tends to produce better outcomes. My own hands-on assessment is in the Rork production review.
A sensible first step
Start by deciding on Android and web, not on which app to build. That one answer picks your product for you.
Then, before spending free prompts, narrow your idea down to three screens. Generate, download the project, and confirm npx expo start runs locally. Getting that far is enough to tell whether this fits the way you work.
For a complete first build, see building your first app with Rork. If you have not settled on native versus web at all, choosing between a native app and a PWA covers the decision one level up.
Pricing, platform coverage, and funding figures all change. Please confirm the current details on rork.com before you commit.
If there is one app you have been meaning to build, it is worth spending those five prompts to find out. Thank you for reading.