Implementing Natural Language Processing in a Mobile Application

Your mobile app processes user reviews, support chats, or document scans. Sooner or later you need to add sentiment analysis, entity recognition, or automatic translation. We've been integrating NLP into mobile apps for five years. Our engineers are certified by Apple and Google, and every project starts with an audit of your data. Whether you're diving into mobile NLP development or comparing online vs offline NLP, this article provides practical advice. Choosing the right NLP architecture determines not only accuracy but also response time, app size, and user privacy. In this article, we share practical experience: how to avoid mistakes and save time and money.

Which Approach to Choose for NLP in a Mobile Application?

The choice between on-device and cloud NLP depends on the scenario. On-device is faster, more private, works offline. Cloud is more accurate, supports more languages, requires internet. We help you pick the optimal solution for your task.

Comparison: On-device NLP processes text 5–10 times faster than cloud provided the model is optimized for mobile. Our practice confirms this.

Mobile NLP Platform APIs: iOS Tokenization and NaturalLanguage Framework

Start with what's already on the device — it's resource-free and works offline.

iOS — NaturalLanguage framework:

import NaturalLanguage

// Language identification
let recognizer = NLLanguageRecognizer()
recognizer.processString("Hello, how are you?")
let language = recognizer.dominantLanguage // .english

// Tokenization
let tokenizer = NLTokenizer(unit: .word)
tokenizer.string = text
tokenizer.enumerateTokens(in: text.startIndex..<text.endIndex) { range, _ in
    print(String(text[range]))
    return true
}

// Sentiment analysis
let tagger = NLTagger(tagSchemes: [.sentimentScore])
tagger.string = text
let (sentiment, _) = tagger.tag(at: text.startIndex,
                                  unit: .paragraph,
                                  scheme: .sentimentScore)
let score = Double(sentiment?.rawValue ?? "0") ?? 0.0
// score: -1.0 (negative) ... +1.0 (positive)

Android — ML Kit Text APIs:

EntityExtraction (ML Kit) can find addresses, phones, dates, tracking numbers, money — offline, model downloads once (~8 MB). LanguageIdentification — analogous to NLLanguageRecognizer. The SmartReply model is covered in another service.

Apple NaturalLanguage Framework and ML Kit for Android are proven tools we use in every second project.

When is Cloud API Better than On-Device? Online vs Offline NLP Comparison

If you need support for rare languages or maximum accuracy (e.g., legal documents), cloud services (OpenAI, YandexGPT, Google Cloud NLP) deliver better quality. We connect them as fallback or main engine when speed requirements are not critical.

Text Classification with TFLite for Mobile NLP

Platform APIs don't handle domain classification — they won't say "this is a food review" or "this is a technical support request." You need your own model.

Typical pipeline for a mobile NLP classifier:

1. Training: BERT-tiny or MobileBERT (8 MB vs 100 MB for full BERT) on PyTorch/TF
2. Conversion: torch.onnx.export() → ONNX → onnxruntime-mobile, or tf2tflite → .tflite
3. Quantization: int8 via TFLite Converter gives ~4x compression with minimal accuracy loss

class TextClassifier(context: Context) {
    private val interpreter: Interpreter
    private val tokenizer: BertTokenizer

    init {
        val modelBuffer = loadModelFile(context, "bert_tiny_classifier.tflite")
        interpreter = Interpreter(modelBuffer, Interpreter.Options().apply {
            addDelegate(NnApiDelegate()) // Android Neural Networks API
        })
        tokenizer = BertTokenizer.fromAssets(context, "vocab.txt")
    }

    fun classify(text: String): ClassificationResult {
        val tokens = tokenizer.encode(text, maxLength = 128, truncate = true)
        val inputIds = Array(1) { tokens.inputIds.toIntArray() }
        val attentionMask = Array(1) { tokens.attentionMask.toIntArray() }
        val output = Array(1) { FloatArray(NUM_LABELS) }

        interpreter.runForMultipleInputsOutputs(
            arrayOf(inputIds, attentionMask),
            mapOf(0 to output)
        )
        return output[0].argmax().let { ClassificationResult(label = LABELS[it], confidence = output[0][it]) }
    }
}

NNAPI delegate on Android 8.1+ speeds up inference through DSP/NPU. On Pixel 7+ — up to 10x acceleration. On budget devices NNAPI may be slower than CPU — test on real hardware.

Named Entity Recognition

NER extracts named entities from text (persons, organizations, locations, dates). Use cases: auto-creating calendar events from messages, pre-filling forms from CVs, parsing receipts.

ML Kit EntityExtraction covers standard entities without training. For custom domains — own NER model based on BiLSTM+CRF or BERT. Model size: BiLSTM — 5–15 MB, DistilBERT-NER — ~60 MB in fp16.

On iOS NL framework returns NLTag with types: .personalName, .placeName, .organizationName. Works on-device offline.

Summarization and Translation

On-device summarization requires heavy models (BART, T5 at least 60–100 MB). For mobile — either extractive summarization (key sentence selection via TF-IDF + MMR, ~100 KB of logic) or cloud API (OpenAI, YandexGPT).

Machine translation: ML Kit Translation supports 59 languages, models download on demand (~30 MB per language pair). On iOS — MLTranslation via Apple Intelligence (iOS 18+) or cloud APIs.

User Input Processing: Cleaning and Normalization

A frequently overlooked step — text preprocessing before the NLP model. Typos, slang, emojis, mixed keyboard layouts (e.g., typing Russian words in Latin letters) all reduce accuracy. Apache Lucene Analyzers on Android or NLTokenizer with custom rules on iOS help normalize the text.

What Our Work Includes

We ensure a transparent process and deliver results with these steps:

1. Analysis of your data and NLP requirements (accuracy, speed, languages)
2. Selection of approach: platform APIs, custom on-device model, or cloud service
3. Model development and training (if needed)
4. Integration into the app with clean architecture (MVVM, MVI)
5. Testing on real devices (30+ iOS/Android models)
6. Documentation and training for your team
7. Post-launch support (3-month warranty period)

Deliverables: API documentation, model access, team training, and 3-month support. Typical investment for platform APIs starts at $5,000, while custom model integration ranges from $15,000 to $50,000, delivering an average 40% time savings.

Why Trust Us with NLP Integration?

• 5+ years in mobile development
• 30+ projects with NLP (from review analysis to voice assistants)
• Certified engineers (iOS, Android, ML)
• Solutions run on 500+ million devices
• Time-to-market acceleration by 2x on average thanks to reusable components
• Our solutions have saved clients an average of 35% on development time

We will evaluate your project in 1 day. Contact us — we'll discuss your task and propose the optimal solution. Order NLP integration today.

Timeline Estimates

Integration of platform NLP APIs (sentiment, tokenization, NER) — 3–5 days. Custom classifier with training and mobile optimization — 3–6 weeks. Timelines are for turnkey work.

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.