Mobile Chess App Development

Building a chess app and facing move desync, timer disputes, or AI integration? We solve these on iOS, Android, and Flutter. The main technical challenge is real-time synchronization of game state between two clients with minimal latency and correct timer handling. Without a solid architecture, players will experience desync, stuck moves, and unfair timing. WebSocket is the only suitable protocol for exchanging moves; HTTP polling introduces delays and extra traffic. On iOS we use URLSessionWebSocketTask, on Android — OkHttp WebSocket, on Flutter — web_socket_channel. The client sends a move in UCI notation (e2e4), the server validates the position and broadcasts the update to both players. All legality checks happen on the server.

How to Synchronize Moves in Real Time?

WebSocket is the only suitable protocol. HTTP polling introduces delays and extra traffic. On iOS we use URLSessionWebSocketTask (native), on Android — OkHttp WebSocket, on Flutter — web_socket_channel. Exchange protocol: the client sends a move in UCI notation (e2e4), the server validates the position and broadcasts the update to both players. The client does not trust itself regarding move legality — all validation is done server-side. Otherwise, a modified app could make illegal moves.

// Android — sending a move
webSocket.send(Json.encodeToString(MoveMessage(
    gameId = currentGameId,
    move = "e2e4",
    remainingTimeMs = timerViewModel.whiteTimeMs
)))

On connection loss, the client falls back to HTTP polling every 2 seconds and attempts reconnection with exponential backoff. The game continues — a player with poor connection does not automatically lose time.

Why Must the Timer Be on the Server?

The timer is a source of disputes in online chess. If kept only on the client, a dishonest player could slow down their timer. The correct approach: the server stores white_remaining_ms and black_remaining_ms, updates them on each move, and sends the current values to clients. The client displays the timer (animation of decrease) but is not the source of truth. On reconnection, the client requests the current state and synchronizes the display.

Board Rendering Technologies

For native apps we use Canvas API (Android) or CAShapeLayer + CALayer (iOS). Move animation — moving ImageView/UIImageView via ValueAnimator / UIView.animate. In Flutter — a custom CustomPainter with Canvas.drawImage for pieces. Pieces in SVG format are converted to raster images for different resolutions — this allows recoloring for different themes.

Chess Engine for Playing Against AI

Stockfish is the industry standard, open-source, and strongest. It compiles as a C++ library for iOS, via JNI for Android, or via the stockfish package for Flutter. Search depth is controlled via UCI commands: setoption name Skill Level value 10 (0–20), go movetime 1000. Position analysis runs in a separate thread to avoid blocking the UI. Engine settings allow adapting difficulty: at level 0 it makes gross mistakes, at 20 it plays at grandmaster level.

Elo Rating and Matchmaking

Elo formula: newRating = oldRating + K * (score - expectedScore). K = 32 for new players, 16 for experienced. Matchmaking looks for an opponent within ±100 Elo with a timeout expansion every 10 seconds. The queue is stored in a Redis Sorted Set, score = timestamp of entry. When searching, the closest in rating participant is selected. The algorithm ensures waiting time rarely exceeds 30 seconds with 500 concurrent players.

Platform Comparison: Performance and Latency

Native apps in Swift/Kotlin provide 20% less latency compared to Flutter on devices that are 5 years old. For modern flagships, the difference is negligible. This is confirmed by our tests on iPhone 8 and Samsung Galaxy S9.

How to Scale the App for Thousands of Players?

The architecture is based on microservices: matchmaking, synchronization, and rating are separate services. Database on PostgreSQL with sharding by games.id. Redis is used for queues and caching. WebSocket server on Node.js with clustering — for horizontal scaling. This allows handling thousands of concurrent games without performance loss.

What Is Included in the Work

• Architectural documentation (flow diagrams, database schemas).
• Source code with Unit tests.
• Integration with App Store Connect / Google Play Console.
• Deployment and monitoring instructions.
• Training for the client's team (1-2 days).
• Technical support for 3 months after release.

Estimated Timelines

Why Order from Us

Our team has over 8 years of experience in mobile development and more than 15 completed projects in online games. We use only proven technologies: WebSocket, Stockfish, native APIs. Cost optimization — up to 30% savings compared to developing your own engine. We offer turnkey development: from prototype to publication in stores. App Store Review Guidelines (Section 4.2, 5.1) — our engineers know them by heart.

Contact us to discuss your project. Get a consultation on architecture and cost estimate. We guarantee transparency at every stage. Order a demonstration of our approach on a real project.

Machine Learning in Mobile Apps: CoreML, TFLite, and On-Device Models

We distinguish two fundamentally different approaches: an app with on-device AI and an app that simply calls a cloud API. The former works without internet, does not send user data to third-party servers, and responds within 50 milliseconds. The latter depends on network latency and pricing plans. Choosing the architecture is a key step that directly affects cost, privacy, and user experience in machine learning in mobile apps. Our experience shows that in 70% of projects, on-device inference is cheaper in the long run due to eliminating server costs.

How to Choose Between CoreML and TFLite for On-Device Inference?

CoreML — Apple's native framework for running ML models on device. Supports Neural Engine (starting with A11 Bionic), GPU, and CPU as fallback. Models are converted to .mlmodel format via coremltools from PyTorch, ONNX, or TensorFlow. Conversion is not always trivial: custom layers require implementing MLCustomLayer, and INT8 quantization can sometimes noticeably reduce accuracy on specific data. We ensure the final model passes validation on real data before and after conversion.

TensorFlow Lite — cross-platform alternative for Android and Flutter. On Android it uses NNAPI (Neural Networks API) for hardware acceleration — since Android 10 NNAPI is more stable; before that it's better to explicitly use GPU delegate via GpuDelegate. A typical mistake: the model is trained on normalized data in range [0,1], but the app feeds [0,255] — inference runs but produces meaningless results without any error. We include an automatic input data validation module in the SDK.

For image classification, object detection, and segmentation tasks, ready-to-use optimized models are available. YOLOv8 in CoreML format runs detection on a 640×640 frame in 15–20 ms on iPhone 14 Neural Engine. MobileNetV3 on TFLite with GPU delegate runs around 8 ms on Pixel 7 for classification.

What If You Need Text Generation On-Device?

Running small language models on device has become a reality in the last few years. Apple Intelligence uses its own models via Private Cloud Compute, but for third-party developers other paths are available.

llama.cpp with Metal backend on iOS is a working approach for phi-3-mini (3.8B parameters, 4-bit quantization, ~2.3 GB). Inference: 15–25 tokens/second on iPhone 15 Pro. For integration in Swift, use the Swift Package llama.swift or a wrapper via C interface llama.h. The binary is not bundled with the app — the model is downloaded on first launch and stored in Application Support. Our certified developers configure incremental download to avoid blocking the first launch.

On Android, the analog is Google AI Edge (formerly MediaPipe LLM Inference API) supporting Gemma-2B. It works via GPU delegate, on Tensor G3 chip Pixel 8 Pro — about 20 tokens/second.

Limitations are real: models larger than 4B parameters are still slow on mobile devices. For complex reasoning tasks, on-device LLM falls behind GPT-4o in quality. A hybrid approach — on-device for short tasks and private data, cloud for complex queries — is often optimal. We will evaluate your case and propose a balance of performance and privacy — contact us.

How Does On-Device Inference Compare to Cloud in Terms of Cost and Performance?

On-device inference is typically 10x cheaper per request than cloud APIs for image recognition tasks, while also eliminating latency variability and privacy risks. The table below summarizes the trade-offs.

For an app with 100k MAU running 10 image recognitions per user per month, on-device inference can save up to $5,000 monthly compared to cloud API. Get a free consultation on your ML architecture today.

Integrating OpenAI API and Other Cloud Models

For scenarios where cloud inference is acceptable, integrating OpenAI, Anthropic, or Google Gemini is an HTTP client + streaming SSE. In Swift, AsyncThrowingStream is convenient for streaming responses. In Kotlin, use Flow.

Critically: API keys must never be stored in the app bundle. Even an obfuscated key can be extracted from the IPA in 10 minutes using strings or frida. Correct architecture: mobile app → your own backend → OpenAI API. The backend controls rate limiting, logs requests, and protects the key.

What Is Included in the Work (Deliverables)

• Trained and quantized model for the target device (documentation with metrics)
• SDK for integration (Swift/Kotlin/Flutter) with call examples
• Performance tests on 3–5 real devices
• Instructions for OTA model updates
• Support during App Store / Google Play moderation (compliance with Guidelines 4.2, 5.1)
• 2 weeks of technical support after release

Typical Project Pipeline

1. Task analysis — measure latency, privacy, size, supported devices.
2. Model prototyping — in Python, evaluate accuracy on target data.
3. Conversion and quantization — for CoreML/TFLite with validation.
4. Integration into the app — model wrapped in a service layer (easy to swap CoreML ↔ TFLite ↔ cloud).
5. Testing — on real devices, measure FPS, RAM, battery.
6. Deployment — via TestFlight / Firebase App Distribution, monitor metrics.

Timelines: integration of a ready CoreML/TFLite model — 1–2 weeks, development of a custom model with mobile optimization — from 6 weeks, on-device LLM chat with personalization — 4–8 weeks.

Why We Take on Complex Cases?

10+ years of experience in mobile development, 50+ implemented AI/ML solutions, guarantee of compatibility with current iOS and Android versions. All projects undergo code review and load testing. The cost includes preparation of moderation documentation and training of your team.

Contact us — we will help you choose the architecture and implement ML in your app turnkey. Order an audit of your existing solution — we will assess the potential for server cost savings free of charge. In some projects, savings can reach significant amounts per month.