Anthropic Claude + RevenueCat + Supabase in a React Native boilerplate

Does AI Mobile Launcher support this combo?

Partially. The AI Pro tier ships OpenAI and Gemini as the two cloud LLM providers out of the box. Anthropic Claude is not bundled today, but the REST client pattern at src/features/gen-ui/api/openai.api.ts is ~80 lines, and porting it to Anthropic is a one-evening job. RevenueCat and Supabase ship as wired. This page is honest about what is shipped vs what you wire.

The stack

React Native 0.83.6 + Expo SDK 55.0.17
TypeScript strict mode
Anthropic: you add a REST client (~80 lines). Recommended models: claude-sonnet-4-6 for quality, claude-haiku-4-5 for cost
RevenueCat: pre-wired with paywall, free trial, restore purchases
Supabase: auth (Apple / Google / email) + Postgres + RLS, tokens in expo-secure-store
State: Redux Toolkit + RTK Query

Setup in five steps

1. Clone the AI Pro tier

git clone <ai-pro-tier-repo>
cd ai-mobile-launcher
pnpm install

2. Add an Anthropic REST client

// src/features/gen-ui/api/anthropic.api.ts
// Mirror the openai.api.ts pattern. POST to:
// https://api.anthropic.com/v1/messages
// Headers: x-api-key, anthropic-version: 2023-06-01
// Then add 'anthropic' to LlmProvider type in store/llm-preferences-slice.ts

# .env
ANTHROPIC_API_KEY=sk-ant-...
ANTHROPIC_MODEL=claude-sonnet-4-6  # or claude-haiku-4-5

3. Supabase env

# .env
EXPO_PUBLIC_SUPABASE_URL=https://<project>.supabase.co
EXPO_PUBLIC_SUPABASE_ANON_KEY=<anon-key>

4. RevenueCat env

# .env
EXPO_PUBLIC_REVENUECAT_IOS_KEY=appl_...
EXPO_PUBLIC_REVENUECAT_ANDROID_KEY=goog_...

5. Dev build

eas build --profile development --platform ios

Why this combo works

Claude wins on two specific kinds of mobile features: long-context summarization (notes, transcripts, document Q&A) and tool-use heavy agents. The 200K context window means you can dump an entire conversation history without pagination tricks, and the tool-use API has been more stable than OpenAI's through 2025-2026.

Pairing it with RevenueCat covers the mobile-specific reality that App Store and Play Store receipt validation is harder than any LLM integration. RevenueCat hides the difference and gives you offer codes, paywall A/B, and revenue analytics out of the box.

Supabase pairs cleanly because pgvector ships in every project, so RAG with Claude as the generator and Supabase as the retriever is a single backend, not two.

What it costs at scale

Line item	1K MAU	100K MAU
Anthropic Claude Haiku (10 chats, 500 tokens)	~$5	~$500
RevenueCat (free up to $2.5K MTR)	$0	~$200
Supabase Pro	$25	~$125
Total infra (excl. store fees)	~$30	~$825

Claude Sonnet costs ~5x Haiku. If your feature is summarization-heavy, route to Haiku by default and only escalate to Sonnet for the long-context calls.

What this combo does NOT cover

Out-of-the-box wiring: Anthropic is not in the AI Pro tier today. You add the REST client
Image generation: Claude does not generate images
Offline / on-device fallback: use llama.rn on-device if you need it
Voice transcription: handled by expo-speech-recognition (native), not by Claude
Streaming responses (SSE): Anthropic supports it but you wire the React Native side

Get this combo

Easiest path: Standard tier ($99), then wire Anthropic yourself (~2 hours). The AI Pro tier ($199) gives you the OpenAI + Gemini REST clients to mirror.

Standard tier ($99)AI Pro tier ($199)

Related combos

OpenAI + RevenueCat + Supabase

Ships in AI Pro out of the box.

Anthropic + RevenueCat + Firebase

Same LLM, Firebase backend instead.