AI Mobile Launcher vs Ignite vs create-expo-app: What $99 Actually Gets You

Short version, because that is what you came for. If you want the best React Native boilerplate for a side project with no deadline, use Ignite. It is free, mature, and Infinite Red has been refining it since the React Native 0.4 era. If you want zero opinions and the official path, use create-expo-app. If you want to ship a paid mobile product in a week with Claude Code already wired in, RevenueCat already pre-configured, and a design system that does not look like a Tailwind starter, AI Mobile Launcher Standard at $99 is the one I would buy. The rest of this article is the honest version of how I got there.

I have shipped on Ignite at one job, used create-expo-app for ~30 prototypes, and now build AI Mobile Launcher full-time. The comparison is not theoretical. It is the same three options I have to pick between every time a client asks me to start a new RN project.

What each option actually is

create-expo-app

Official starting point from the Expo team. Runs npx create-expo-app@latest. Drops you into a typed Expo project with a default Tabs template, expo-router preconfigured, and almost nothing else. Free. Maintained directly by the platform owner.

Best for: prototypes, learning the platform, or building something where you genuinely want to make every architectural decision yourself.

Ignite (Infinite Red)

The most-starred React Native boilerplate on GitHub. Opinionated stack, MobX-State-Tree (or you can swap), Reactotron integration, Ignite CLI for generating screens and models. Maintained by Infinite Red. Mature, free, MIT licensed.

Best for: teams that want a well-known opinionated base, do not need AI patterns wired in, and have engineering time to make the stack their own.

AI Mobile Launcher

Three tiers. Lite is free on GitHub (limited features, MIT). Standard is $99 and ships the full RN production stack with Supabase auth, RevenueCat paywall, RTK Query data layer, MMKV cache, Restyle theme with type-safe tokens, and Palette B (~60 components). AI Pro is $199 and adds Dynamic Onboarding via Wire RN, on-device LLM via llama.rn, and cloud LLM routing.

The wedge: every tier ships with U-AMOS pre-wired, the SDD (Spec-Driven Development) pipeline as 5 Claude Code sub-agents, and an opinionated set of skills that make Claude productive on day 1.

Best for: solo devs and small teams shipping paid mobile products. Indie hackers who want to launch in 1-2 weeks. Senior RN engineers who already know they want RevenueCat + Supabase + RTK Query and do not want to wire it for the 12th time.

The feature matrix

• Price. create-expo-app: free. Ignite: free. AIM-L Standard: $99 one-time. AIM-L AI Pro: $199 one-time.
• State management. create-expo-app: none. Ignite: MobX-State-Tree. AIM-L: Redux Toolkit + RTK Query.
• Auth. create-expo-app: none. Ignite: none. AIM-L: Supabase + Apple/Google Sign-In.
• Paywall / IAP. create-expo-app: none. Ignite: none. AIM-L: RevenueCat wired.
• Design system. create-expo-app: none. Ignite: basic theme. AIM-L: Palette B (~60 components).
• Storage. create-expo-app: AsyncStorage. Ignite: MMKV. AIM-L: MMKV encrypted.
• Claude Code skills. create-expo-app: none. Ignite: none. AIM-L: U-AMOS + SDD pipeline (5 sub-agents).
• AI features. create-expo-app: none. Ignite: none. AIM-L AI Pro: local LLM (Phi-3) + cloud routing + streaming.
• Maintenance. create-expo-app: Expo team. Ignite: Infinite Red. AIM-L: indie (me).

A few caveats on the matrix. AIM-L's “commercial use license” means you can ship as many of your own apps as you want. It is not a reseller license. The “indie maintained” row matters. Ignite has Infinite Red behind it. AIM-L is one engineer. That is a real tradeoff. I update it weekly because I dogfood it on Morrow Self and on every client project, but my bus factor is 1 and Infinite Red's is not.

Where Ignite wins

• Brand age + community. 8+ years of usage, Stack Overflow answers exist, the patterns are well-documented externally. Hiring a contractor who already knows the codebase is realistic.
• OSS license + no money. Free is free. For a side project that may not ship, free is the right answer.
• Reactotron integration. Excellent dev tooling, especially for MobX-heavy apps.
• Generators. ignite generate screen Foo is a solid CLI. Faster than typing the boilerplate.

If you want a battle-tested OSS base and your timeline is in months, not days, Ignite is the boring correct answer.

Where create-expo-app wins

• Officialness. Expo team maintains it. Breaking changes get addressed before SDK release.
• Minimal opinions. You can build literally anything on top. Nothing to undo.
• Free. Same as Ignite.
• Learning material. When you are still learning RN, the smaller surface is easier to read.

If you are learning RN, prototyping a new idea every weekend, or building something genuinely unusual where the opinions of others would be in the way, this is correct.

Where AI Mobile Launcher wins

• Time to first paid user. Supabase auth + RevenueCat + Stripe wired. You can charge a customer on day 3 of a new project. With Ignite or create-expo-app, that wiring is 3-5 days of work.
• U-AMOS pre-wired. Claude Code is productive on day 1. Memory bank, context map, anti-hallucination checklist, plan/act mode. The CLAUDE.md already exists, indexed.
• The SDD pipeline. 5 Claude Code sub-agents (researcher, planner, implementer, tester, reviewer) and 8 slash commands. AI dev workflow built into the boilerplate, not bolted on later.
• Palette B design system. ~60 components, Restyle-typed tokens, the kind of UI that does not scream “Tailwind starter.” You get to focus on your product surface, not on spacing rules.
• AI Pro upgrade path. When you want to add Dynamic Onboarding, on-device LLM, or cloud LLM routing, you upgrade tier. You do not rebuild.

The wedge in one sentence: AIM-L is the only one of the three with AI workflows already wired into the boilerplate. Ignite is a great RN starter. AIM-L is an AI-RN starter. Different category, slightly different audience.

When I would still not recommend AIM-L

This is the part most “vs” articles skip. Buying my product is sometimes the wrong call.

• You hate opinions. If you want to pick your own state library, ORM, auth provider, design system, you are going to spend half a day removing my choices. Use create-expo-app.
• Your project is exploratory. If you are not sure you will ship it, $99 is fine but free is better. Use Ignite or create-expo-app.
• Your team has a strong Ignite muscle already. Switching costs are real. If your senior engineers already pattern-match on Ignite, the productivity loss from re-learning my conventions is not worth the AI-workflow gain.
• You do not use Claude Code. A real chunk of the $99 value is the U-AMOS + SDD wiring. If you are a Cursor-only or Copilot-only shop, you still get the RN stack and the design system, but you are paying for things you will not use.

If two or more of those are true, skip AIM-L. The honest math beats the sales math.

What “best React Native boilerplate” actually means

The reason there are 40 boilerplates on the market is that “best” is context-dependent. The framing I find most useful:

• Best for free + community + OSS: Ignite.
• Best for official + minimal + free: create-expo-app.
• Best for shipping a paid AI mobile product in 1-2 weeks: AI Mobile Launcher Standard.
• Best for shipping an AI-native mobile product with on-device LLM and streaming: AI Mobile Launcher AI Pro.

These are not competing on the same axis. Picking is mostly about what you are trying to ship and how soon.

The 5-minute decision tree

• Building a side project with no deadline? Use Ignite.
• Learning RN or building something unusual? Use create-expo-app.
• Shipping a paid mobile product in 1-2 weeks, want Claude Code productive on day 1? Use AIM-L Standard.
• Building an AI-native mobile product (on-device LLM, streaming chat, agentic flows)? Use AIM-L AI Pro.

The Lite tier is on GitHub if you want to read the structure first. Standard and AI Pro are at aimobilelauncher.com/expo-boilerplate and /ai-expo. One issue per week on AI-native mobile development at codemeetai.substack.com.

FAQ

Is AI Mobile Launcher worth $99 over Ignite?

If you use Claude Code and want to ship a paid product in days instead of weeks, yes. The U-AMOS skeleton, SDD pipeline, RevenueCat wiring, and Palette B design system save 4-8 days of work, even before the AI features. If you do not use Claude Code and have months to spend, Ignite is the better choice.

Why not just fork Ignite and add the missing pieces?

You can. Plan on 5-10 days of work to wire Supabase auth, RevenueCat, RTK Query, a design system, and Claude Code skills onto an Ignite base. If your day rate is anywhere above $20, the math points to buying. If your time is unconstrained, forking is reasonable.

Does AIM-L work without Claude Code?

Yes. The boilerplate is a full RN production stack and runs without any AI tooling. You lose the U-AMOS + SDD productivity layer if you do not use Claude Code, but the RN scaffolding still saves you the auth + payments + design wiring work. The AI Pro features (Dynamic Onboarding, local LLM, streaming) work at runtime independent of the dev tooling.

What is the difference between AIM-L Standard and AI Pro?

Standard ($99) is the full RN production stack with no AI features. AI Pro ($199) adds Dynamic Onboarding via Wire RN, on-device LLM via llama.rn (Phi-3 Mini), cloud LLM routing across OpenAI / Anthropic / Gemini, and Hermes-safe streaming. AI Pro is for products where AI itself is part of the product surface.

Is AIM-L a good React Native ignite alternative for AI apps?

For AI-native mobile apps, yes. Ignite is a strong general-purpose RN boilerplate but does not ship AI patterns wired in. AIM-L Standard wires the Claude Code productivity layer; AI Pro wires the on-device and cloud AI patterns. If your product needs an LLM-driven UI, AIM-L is the closer fit.

Related reading

• Best React Native boilerplate 2026
• AIM-L vs Ignite vs create-expo-app: an honest comparison
• Vibe coding boilerplate: why blank Expo projects break Cursor