# x1 launches an AI iPhone builder that asks questions before writing code > Source: > Published: 2026-08-26 18:31:29+00:00 # x1 launches an AI iPhone builder that asks questions before writing code **Manil Lakabi's YC-backed tool plans, designs and packages React Native apps for testing and App Store review.** By [Ryan Merket](/author/ryan-merket) · Published · Updated Primary source: [Product Hunt](https://www.producthunt.com/products/x1-2?utm_source=other&utm_medium=social&bc=1&page=5#comments) ## Why it matters AI app builders have made prototypes cheap. x1 is betting that planning, project memory and App Store preparation will become the valuable layer when users try to ship and maintain them. [Manil Lakabi (@lakabimanil_)](https://x.com/lakabimanil_), founder of [x1](https://x1.new/), launched an AI app builder on August 26th that guides users from a rough idea through design, development and App Store submission instead of attempting to generate an entire iPhone app from one prompt. The San Francisco-based founder is making a process bet in a category built around speed. In an [August 26th product listing](https://www.producthunt.com/products/x1-2?utm_source=other&utm_medium=social&bc=1&page=5#comments), the x1 app builder described a workflow that asks focused questions, maps the product into milestones, lets the user review each design and then builds the app in stages. Users can preview the result on an iPhone as they go. When the app is ready, x1 prepares screenshots, listing copy and a submission through the user's Apple Developer account. Lakabi previously worked on the generative AI team at Scale AI and the Facebook Reality Labs team at Meta, according to [Y Combinator's profile of x1](https://www.ycombinator.com/companies/x1). He founded x1 in 2024 and took it through YC's Fall 2024 batch. YC lists x1 as a five-person team. ### Planning before prompting x1 uses the pitch "Lovable for iPhone apps," borrowing the web-app builder's category recognition while rejecting the one-prompt interaction model associated with many AI builders. Lakabi's product separates app creation into decisions about the idea, design, build, publishing and revenue model. The distinction reflects Lakabi's thesis that AI-generated software becomes harder to maintain as users revise it. A convincing first screen is easy to produce. Authentication, payments, permissions, stored data and later feature changes expose assumptions that the model made earlier in the process. "The prompt box is not the final interface for building software," Lakabi says on [x1's about page](https://x1.new/about). x1 attempts to capture product decisions before generation, then preserve them as the project moves through design and engineering. The [documented workflow](https://x1.new/how-it-works) begins with an idea or uploaded file, followed by a feasibility check. x1 then asks clarifying questions and generates a product plan covering the target user, screen map, core flows, milestones and launch priorities. Users review interfaces in a visual canvas before x1 generates each milestone and runs its own QA pass. Lakabi said the product is not trying to prove that its models write better code than Claude Code, Codex or Replit. Instead, the company is packaging the planning, scoping, context management and verification work around code generation for users who may not know how to perform those jobs themselves. The system is designed to update dependent screens, flows and features when a user changes an earlier decision rather than treating each revision as a new prompt. That process gives x1 a sharper position than a general-purpose coding assistant. The tradeoff is scope. The company is starting with iPhone apps rather than supporting Android and other platforms at launch. Lakabi said the choice reflects both a desire to concentrate on iOS design, subscriptions, testing and App Store submission, and the team's view that iPhone users offer app builders a stronger path to monetization. Android support is planned, but the company gave no timetable. x1's own [examples page](https://x1.new/examples) says complex 3D games, Android-first products, heavily regulated apps, custom hardware projects and products requiring advanced integrations are poor starting points. The displayed portfolio consists of demo concepts rather than customer case studies. ### React Native, credits and a closed workflow x1's current product specification says it generates React Native apps that users can preview on an iPhone and distribute through Apple's release workflow. The company also says the builder can incorporate payments, subscriptions, databases, authentication, AI services and third-party APIs. Muhammad Inamullah, a member of the launch team, said Supabase and OpenAI services can be configured within x1, while an integration agent can set up other providers and connect their credentials. Those integration claims have not been independently tested, and users cannot yet export a generated app to continue development elsewhere. Inamullah said source export is under development but "not too far out." For now, building, testing and publishing remain inside x1, although the finished app is submitted through the user's own Apple Developer account. New accounts receive 100 credits without a credit card. x1 says that allowance can generally cover idea validation, a product plan, several design explorations and one built milestone. Every AI action consumes credits, including planning, design generation, building and QA. Monthly plans cost [$20, $50 and $100](https://x1.new/pricing), providing 1,000, 2,500 and 5,000 credits, respectively. Annual billing lowers the effective monthly prices to $8, $20 and $40. Publishing to TestFlight and the App Store starts with the paid Starter plan. The credit model makes the guided workflow part of x1's cost control. Decisions caught during planning should require fewer generated revisions later. Whether that holds across complicated products will depend on how consistently x1's feasibility checks, project memory and milestone QA prevent users from entering the same prompt-and-fix loops the product was designed to replace. ### Apple still decides what ships x1 can prepare a submission and move files through the release workflow. Apple retains control of distribution. Developers need a separate [Apple Developer Program membership](https://developer.apple.com/programs/), which costs $99 per year in the United States, and every submitted app remains subject to [Apple's review process](https://developer.apple.com/app-store/review/). That boundary matters because "idea to App Store" compresses several distinct jobs into one phrase. x1 can generate the application, organize launch materials and help submit it. Apple can still reject the result for technical, privacy, safety, content or design problems. The builder also cannot supply demand, retention or a viable business model for the app it produces. Lakabi's narrower claim is the consequential one: a solo builder should be able to coordinate product planning, design, engineering and release work in one system. x1 is using iOS as the test case, where a single distribution channel and a strict review process make unfinished software difficult to disguise. If its staged workflow works, the advantage will come from carrying decisions through the entire build, long after the first generated screen stops looking impressive.