AI Models
Hands-on guides for integrating Gemini, Claude, and GPT into Rork apps, along with prompt design techniques and in-depth AI app builder comparisons.
Is Your Rork App Still Calling Sonnet 4.5? Finding Out Before November 30
Claude Sonnet 4.5 is scheduled to retire from the API on November 30, 2026. A small script to find leftover model IDs in your app, and the order to fix them in depending on where they live.
Rork's Design Mode or Plan Mode? Sort It by What You Are Unsure About: the Look or the Scope
How to choose between Build, Plan, and Design in Rork by what you are unsure about. Three questions, three examples, what the free plan can settle, and how to write a first message.
When Rork's Fix Breaks the Screen Next Door: Choosing Between Undo, Revert, and Edit Message
When an AI fix in Rork backfires, three questions tell you whether to Undo, Revert, or Edit message. Going back costs no credits; here is what it cannot bring back, and how to ask again.
Before You Type "Connect Supabase": The Two Values to Copy, and the One You Must Not
Rork asks for just two things before you type Connect Supabase in chat: a Publishable key and your project link. Here is where to copy them from, how to avoid pasting the secret key that sits on the same screen, and a small script that checks the values locally first.
Before You Ship an AI-Made App Icon or Image: A Ledger and Four Checks
Can you ship an AI-generated icon or illustration as is? Before deciding, run four checks: a provenance ledger, the tool's terms, resemblance to existing work, and store size rules, with a working script.
Three checks to run before you ask the AI again about a black preview screen
When a fresh Rork preview opens to a black screen, split the problem into three layers — device, rendering, native — before sending the same fix request again.
Take the Prompt Out of the App: An Edge Function Gateway With the Guard on the Output Side
A system prompt shipped inside your app is readable from the day it ships. This walkthrough moves the prompt into a Supabase Edge Function, lets the client name only a task, and puts the guard on an output allowlist, with tested code.
Sonnet 5.5, GPT-6.1 Sol, or Opus 5.5? I Decide by the Kind of Fix Before I Pick the Model. A Week of Logging Where My Rork Credits Went
After Rork's model menu settled into Sonnet 5.5, GPT-6.1 Sol, and Opus 5.5, I sorted my fixes into three kinds and assigned a model and effort level to each. Here is how I read the changelog, the tiny ledger that tracks where each credit goes, and the rule I drew from a week of entries.
Keep Fixing in the Same Thread, or Restore and Rephrase? What I Learned Fixing One Rork Bug Both Ways
When a Rork fix request fails for the third time, do you keep going or use Restore and ask differently? I fixed the same bug both ways and compared round trips, leftover code, and where I now draw the line.
The recommended replacement was already gone — auditing all 74 rows of Gemini's deprecation table
Cross-checking Gemini's deprecation table turned up five rows where the recommended replacement had itself already shut down, including gemini-2.5-flash-image, which goes away on October 2. Here is how to verify a migration target, and how to catch the 404 before your users do.
Where the AI decision lives once Rork splits your app into two native codebases
New Rork apps are Swift on iPhone and Kotlin on Android — two separate codebases. Here is how I decide which AI-related judgments stay in each client and which move to the shared backend, plus the minimal contract both clients read the same way.
Is That API Key Sitting Inside Your App? Rork Environment Variables vs. Supabase Edge Functions
Rork gives you two places to put an API key: its own environment variables, or a secret inside a Supabase Edge Function. Here is how I decide between them based on what I can revoke and who gets the bill, plus three checks to run after the key is in place.
Ask Rork for an iOS 27 feature and you may get last year's code back, with no error at all
iOS 27 ships on September 14. When a model is asked for an API it has never seen, there are two kinds of failure: the kind that stops your build, and the kind that does not. Here is why standard Rork and Rork Max differ, and the order I check generated code in.
Cutting MCP Tools Didn't Make Anything Lighter — 2,377 Bytes of Definitions vs 110,298 Bytes of Response
I wrote a minimal MCP server for a wallpaper catalog and measured the byte cost of tool definitions against the byte cost of responses. Here is which side actually matters, and the real reason to merge tools.
The Three-Minute Video That Kept Failing — Moving Rork's Gemini Uploads Off the Worker
How I rerouted video uploads to the Gemini Files API from a Cloudflare Workers relay to a direct device-to-Google path — why the 128MB isolate limit breaks the relay, how to hand out a resumable upload URL without leaking your API key, what Expo SDK 54 broke on the device side, and how the 20GB project ceiling and media resolution decide the bill.
On-Device Image Classification: TFLite on React Native or Core ML on Rork Max — How I Chose After Building Both
Adding on-device image classification means choosing between TFLite on React Native and Core ML on Rork Max. I built the same feature both ways, measured the end-to-end breakdown, and worked out what the decision actually hinges on.
When Your Finance App's AI Keeps Dumping Everything Into 'Other' — Field Notes on Catching Silent Classification Drift
An AI expense classifier in a Rork finance app can be accurate at launch and quietly decay over months until the monthly advice goes wrong. Here is how I instrumented confidence and category distribution to get ahead of the drift, with working code.
Integrating Image Playground in Rork Max Native Swift — Availability Design and Fallbacks for In-App Image Generation
How to build Image Playground into a Rork Max Swift app: availability checks with supportsImagePlayground, the imagePlaygroundSheet modifier, programmatic generation with ImageCreator, and fallback design for unsupported devices.
Long-Form On-Device Transcription with SpeechAnalyzer in Rork Max's Native Swift
Implementation notes on rebuilding long-form, offline transcription with iOS 26's SpeechAnalyzer and SpeechTranscriber after hitting the walls of SFSpeechRecognizer. Covers model asset downloads, feeding audio through an AsyncStream, drawing volatile vs. final results, and the boundary design for Rork Max native code and bridging from Expo — with the pitfalls I actually hit.
When Your Rork Hybrid AI Quietly Drifts to the Cloud and the Bill Creeps Up — Field Notes on Instrumenting Routing Decisions
A router that splits work across on-device, edge, and cloud layers will quietly drift toward the cloud when no one logs its decisions — flat traffic, rising bill. These are field notes on instrumenting routing to isolate the cause.
Monetizing a Rork-Built App — Choosing Between Ads, Subscriptions, and Freemium
How to monetize an app built with Rork — from choosing between ads, subscriptions, freemium, and one-time purchase to the implementation details. Phased AdMob formats, treating ad-free as a single source of truth, and price anchoring, written from the indie-developer trenches.
Pushing Rork Max's AI Beyond Vibe Coding — Patterns That Actually Improve Implementation Quality
A deep-dive on using Rork Max as an implementation partner: prompt patterns, context management, choosing between plain Rork and Rork Max, building without burning credits, and verifying AI-written native code before App Store submission — drawn from solo indie experience.
Receiving On-Device AI Output as Typed Data with Foundation Models Guided Generation
How to receive Foundation Models output as typed Swift structs instead of free text, with working code for Guided Generation and Tool Calling on-device.
On-Device Translation in a Rork Max App with iOS 18 — Free, Offline, Multilingual
Add free, offline, real-time translation to a Rork Max Swift app using the iOS 18 Translation framework. Covers checking language availability, batch translation, and avoiding empty results — all with working Swift code.
Rork Max Credits Drain on Rework, Not Features — Where to Let AI Run and Where to Finish by Hand
Rork Max credit usage is driven less by how complex your app is and more by how many times you regenerate a screen. Here is a practical way to decide what to hand to AI and what to finish by hand, from an indie developer who ships to the stores solo.
Before You Pay $200/mo for Rork Max, Map How Far Expo Reaches in Three Tiers
Wanting widgets or Live Activities makes Rork Max tempting, but most of those features are reachable from the Expo setup that standard Rork generates. Here is how I sort each Apple-native feature into three tiers—reachable in Expo, reachable with a custom module, or where Max is the pragmatic answer—and verify which tier my app is in before paying.
Making Credits Add Up in a Rork AI Image App — Field Notes on Atomic Ledgers and Moderation
Credit billing in a Rork AI image and video app breaks in production because of the order between generation and deduction. Here are the field notes — atomic ledger consumption, idempotency, refunds on failure, and moderation — with Supabase code you can ship.
Where to Stop Letting Rork Fix Your Bugs: A Triage Routine for the 30% That Need You
Most bugs you hand Rork get fixed in a couple of regenerations. A stubborn minority loop forever, each fix spawning a new symptom. Here is the triage routine I use to split what to delegate from what to take over by hand, with retreat lines, regression guards, and a decision log.
Building a Room-Measuring App with Rork Max and RoomPlan — Absorbing Scan Variance by Design
Now that Rork Max generates native Swift, a LiDAR-based RoomPlan scanning app is a realistic project for an indie developer. Scan results drift depending on device and how you hold it. This walks through how to absorb that drift in your app design and export stable dimensions.
On-Device Image Tagging in Rork Max Swift Apps with Foundation Models Image Input
WWDC26 gave the on-device Foundation Models model image input. Here is how to add image tagging and captioning to a Rork Max Swift app entirely on-device, including the availability gate, structured output, and Vision interop.
Calling Apple Foundation Models from a Rork (Expo) App: Bridging On-Device AI Through a Native Module
Rork generates Expo (React Native) apps, but Apple Foundation Models ships as a Swift framework you can't touch from JavaScript. Here's how to write an Expo Modules API bridge, gate it by availability, and fall back to the cloud on unsupported devices.
Routing inference on-device first and escaping to the cloud only when it's worth it, in a Rork app
Build a tiered, fallback-based inference router in a Rork (Expo) app: cache to on-device to Private Cloud Compute to a remote API (Claude/Gemini). Working TypeScript covering budgets, timeouts, caching, and image routing.
A Three-Layer AI Cost Design for Rork Apps After Apple Opened Foundation Models to Small Developers
Apple now offers Foundation Models on Private Cloud Compute at no charge for developers under two million first downloads. Here is a three-layer cost architecture for Rork apps, with a simulation script and working bridge code.
What Did Rork Actually Change as an AI Mobile App Builder? A Six-Month Field Review
An honest, indie-developer review of Rork as an AI mobile app builder after six months of real-world use. What changed, where it shines, where it gets stuck, and how to decide whether to start with it.
One Month with Rork Max AI Cloud: Latency, Cost, and the Hybrid Setup I Settled On
A month of running Rork Max AI Cloud beside local M-series execution: p50/p95 latency distributions, real monthly cost, the benchmark script I used, and a budget ledger that actually stops spending.
Picking Rork Max Over FlutterFlow and Replit Agent — Selection Criteria from an Established App Business
I ran Rork Max, FlutterFlow, and Replit Agent in parallel for six weeks while adding a new AI-wallpaper feature to an existing wallpaper app business at Dolice (cumulative ~50M downloads since 2014). Greenfield comparisons are everywhere; this one is from the rarer angle of fitting an AI app builder into an existing app business — and why Rork Max won.
A month of running Rork and Claude on Xcode side by side — where each one ended up fitting
Notes from an indie developer who has shipped wallpaper apps on iOS and Android since 2014. After a month of keeping both Rork and Claude on Xcode open on the same desk, the work that fit each one turned out to be more obvious than I expected.
How I Stop Rork's AI From Generating Outdated Library Code — My Version-Pinned Prompt Template
A practical look at why Rork Max sometimes generates outdated API code, the version-pinned prompt template I rely on every day, and the device-level checks I use to catch the cases where mismatches still slip through.
7 Real Challenges When Building a Tarot App with Rork (And How to Solve Them)
A practical guide to building a daily tarot and oracle card app with Rork. Covers card data architecture, AI interpretation generation, daily reset logic, monetization design, and App Store review pitfalls specific to fortune-telling apps.
Building an AI Writing Coach App with Rork Max — Complete Implementation Guide for Claude API Streaming, RevenueCat Subscriptions, and History Management
A complete guide to building an AI writing coach app using Rork Max and Claude API. Covers backend architecture, SSE streaming, RevenueCat subscription tiers, MMKV history persistence, and App Store strategy — production-ready code throughout.