HealthyCropping – Offline crop disease diagnosis on Android (TFLite and RAG)

HealthyCropping, an open-source, non-commercial Android app published under the package name com.keylorpineda.healthycropping, runs quantized MobileNetV3 TFLite models on-device to diagnose coffee leaf rust, cacao frosty pod rot and banana black Sigatoka in under 50ms without a cellular connection, storing scans in SQLite and syncing to a FastAPI backend with a Gemini RAG advisory layer grounded in ICAFE, CATIE and CORBANA literature when connectivity returns. The project's docs/PLAN.md tracks 9 phases: Phases 1 through 5 (monorepo foundation, coffee model pipeline with INT8/FP16 export and verified 224x224 RGB parity, Expo React Native shell, edge inference and scanner, and cacao and banana specialist models) are completed, while Phase 6 (FastAPI and cloud sync) and Phase 7 (grounded Gemini RAG) are in progress. Phase 8 (low-end Android benchmarks and field daylight testing) and Phase 9 (signed APK distribution and a farmer pilot in Tarrazu, Costa Rica) are scheduled, and the app's disclaimer states it supports field decisions and does not replace a certified agronomist or phytosanitary laboratory.

Smallholder and medium-scale producers of coffee Coffea arabica / canephora , cacao Theobroma cacao , and banana Musa spp. in Central America frequently work in mountainous or rural plots with intermittent or zero cellular connectivity. Identifying pathogens early—such as coffee leaf rust Hemileia vastatrix , cacao frosty pod rot Moniliophthora roreri , or banana black Sigatoka Pseudocercospora fijiensis —is critical to timing cultural controls and preventing yield loss, yet cloud-only diagnostic tools fail in the field when connectivity drops. HealthyCropping com.keylorpineda.healthycropping is an open-source, non-commercial, offline-first Android application. It runs quantized computer vision models directly on the phone with sub-50ms inference, records scans and plot geometries locally in SQLite, and synchronizes with a FastAPI backend and Gemini RAG advisory layer grounded in certified institutional literature ICAFE, CATIE, CORBANA once connectivity is restored. Agronomic Disclaimer : HealthyCropping is a non-commercial, open-source decision-support tool designed initially for Costa Rica. It supports field decisions and does not replace a certified agronomist or phytosanitary laboratory. Always consult an agricultural professional before applying agrochemicals or removing plants. The project is structured into 9 execution phases according to docs/PLAN.md https://github.com/keylorpineda/healthycropping/blob/develop/docs/PLAN.md : | Phase | Milestone | Scope | Status | |---|---|---|---| | Phase 1 | Foundation & Governance | Monorepo layout, strict linting ruff , mypy , ESLint , CI quality gates | Completed | | Phase 2 | Coffee Model Pipeline | MobileNetV3 TFLite, INT8/FP16 export, verified 224x224 RGB parity | Completed | | Phase 3 | Core Mobile Shell | Expo React Native, SQLite schema, agricultural design system, Spanish UI | Completed | | Phase 4 | Edge Inference & Scanner | On-device TFLite camera, reticle framing, confidence thresholding, triage | Completed | | Phase 5 | Multi-Crop Expansion | Cacao pod and banana leaf specialist models, multi-crop app routing | Completed | | Phase 6 | FastAPI & Cloud Sync | Farm/plot sync endpoints, Firebase Admin auth, Cloudinary image ingestion | In Progress | | Phase 7 | Grounded Gemini RAG | Institutional agronomy corpus, citation retrieval, safe treatment advice | In Progress | | Phase 8 | Performance & Hardening | Low-end Android hardware benchmarks, field daylight testing | Scheduled | | Phase 9 | Field Validation & Release | Signed APK distribution, farmer pilot in Tarrazu, Costa Rica | Scheduled | php flowchart LR A "1. Select Crop\n Coffee / Cacao / Banana " -- B "2. Capture Photo\n Guided framing tips " B -- C "3. On-Device Preprocessing\n 224x224 RGB + normalize " C -- D "4. Offline TFLite Inference\n MobileNetV3 float16 / INT8 " D -- E "5. Instant Local Result\n Class, Severity, SQLite " E -. "Optional when online" .- F "6. Cloud Sync & Gemini RAG\n FastAPI + Cited Manuals " 1. Crop Selection : The user selects Coffee leaf underside , Cacao pod , or Banana leaf prior to scanning so a dedicated specialist model runs for that crop. 2. Offline On-Device Inference : The captured photo is resized to 224x224 , decoded to raw RGB, normalized with exact training statistics, and evaluated locally in < 50ms using react-native-fast-tflite . 3. Actionable Triage & Thresholding : If confidence falls below the calibrated validation threshold, the app instructs the user to retake the photo. Otherwise, it reports predicted class, severity level, confidence, and recommended triage action healthy , watch , or act now . 4. Local Persistence & Optional Sync : Scans, farms, and plots are stored locally in expo-sqlite with client-generated UUIDs. Account creation is optional and used solely to synchronize records idempotently through the FastAPI backend api/ . 5. Grounded RAG Advisory : When online, the backend retrieves passages from official technical manuals ICAFE for coffee, CATIE for cacao, CORBANA for banana and utilizes Gemini to summarize recommendations without hallucinated chemical remedies. One fine-tuned MobileNetV3 model is trained, evaluated, and exported per crop 224x224 RGB input, quantized TFLite < 10 MB : | Crop | Target Organ | Target Classes | Evaluation Metric | Export Format | Documentation | |---|---|---|---|---|---| | Coffee | Leaf underside | healthy , leaf rust + severity , leaf miner , cercospora , brown leaf spot , red spider mite | 90% Macro-F1 | FP16 / INT8 TFLite | Model Card https://github.com/keylorpineda/healthycropping/blob/develop/docs/MODEL CARD.md 2-coffee-leaf-model-coffee | | Cacao | Pod | healthy , monilia by stage , black pod Phytophthora | 92% Macro-F1 | FP16 / INT8 TFLite | Model Card https://github.com/keylorpineda/healthycropping/blob/develop/docs/MODEL CARD.md 3-cacao-pod-model-cacao | | Banana | Leaf | healthy , black sigatoka , fusarium wilt , cordana | 92% Macro-F1 | FP16 / INT8 TFLite | Model Card https://github.com/keylorpineda/healthycropping/blob/develop/docs/MODEL CARD.md 4-banana-leaf-model-banana | Tensor preprocessing parity between Python ml/ and React Native app/src/ml/ is enforced by shared fixture tests. healthycropping/ ├── app/ Expo React Native Android, pnpm : on-device TFLite, SQLite, UI ├── api/ Python 3.12 uv : FastAPI, Firebase Admin, Cloudinary, RAG ├── ml/ Python 3.12 uv : data preprocessing, training, eval, TFLite export └── docs/ ARCHITECTURE, PLAN, MODEL CARD, DATASETS, EXPERIMENTS, DEVLOG | Workspace | Stack | Purpose | |---|---|---| | app/ https://github.com/keylorpineda/healthycropping/blob/develop/app/README.md | React Native 0.76, Expo 52, TypeScript, SQLite, TFLite | Offline mobile application for Android | | api/ | Python 3.12, FastAPI, Firebase Admin, Cloudinary, Gemini | Sync endpoints, token verification & RAG advisory | | ml/ | Python 3.12, TensorFlow 2.16, TFLite, Pillow | Training pipelines, evaluation & quantized edge export | - Mobile : Node.js 20+, pnpm configured with node-linker=hoisted , Android SDK. - Backend & ML : Python 3.12, uv package manager. cd app pnpm install pnpm typecheck pnpm lint pnpm test pnpm start cd api uv sync --all-extras uv run ruff check . uv run mypy src tests uv run pytest uv run uvicorn src.main:app --reload cd ml uv sync --all-extras uv run ruff check . uv run mypy src scripts tests uv run pytest Contributions are welcome. Please check our open issues and guidelines before opening a pull request: - Roadmap & Tasks : https://github.com/keylorpineda/healthycropping/issues https://github.com/keylorpineda/healthycropping/issues - Community Spotlight : GitHub Discussions Announcement https://github.com/community/community/discussions/137746 discussioncomment-19118448 - Contributing Guidelines : CONTRIBUTING.md https://github.com/keylorpineda/healthycropping/blob/develop/CONTRIBUTING.md - Architecture Overview : docs/ARCHITECTURE.md https://github.com/keylorpineda/healthycropping/blob/develop/docs/ARCHITECTURE.md - Code of Conduct : CODE OF CONDUCT.md https://github.com/keylorpineda/healthycropping/blob/develop/CODE OF CONDUCT.md 1. Decision Support Only : Visual symptoms cannot distinguish every microscopic pathogen strain or nutrient deficiency. Laboratory diagnostics remain necessary for high-impact interventions. 2. Field Lighting Sensitivity : Direct specular glare on wet foliage or extreme motion blur can reduce model confidence. 3. Backend Cold Starts : The optional FastAPI synchronization API runs on Render's free tier sleeps after 15 minutes of inactivity . Core disease diagnosis and SQLite storage operate 100% offline on the phone and are never affected. The MIT license in LICENSE https://github.com/keylorpineda/healthycropping/blob/develop/LICENSE applies solely to the source code in this repository. Training datasets remain under their respective upstream licenses documented in docs/DATASETS.md https://github.com/keylorpineda/healthycropping/blob/develop/docs/DATASETS.md : - RoCoLe Coffee — Mendeley Data, DOI 10.17632/c5yvn32dzg.2 https://doi.org/10.17632/c5yvn32dzg.2 - BRACOL Coffee — Mendeley Data, DOI 10.17632/yy2k5y8mxg.1 https://doi.org/10.17632/yy2k5y8mxg.1 - CocoaMoniliaDataSet Cacao — Zenodo, DOI 10.5281/zenodo.17156051 https://doi.org/10.5281/zenodo.17156051 , CC BY 4.0 - UNAB Cocoa Diseases Cacao — Kaggle serranosebas/enfermedades-cacao-yolov4 - Banana Leaves Imagery Banana — NM-AIST DSpace 20.500.12479/3003 https://dspace.nm-aist.ac.tz/handle/20.500.12479/3003 ; DOI 10.1038/s41597-025-04456-4 https://doi.org/10.1038/s41597-025-04456-4 - BananaLSD Banana — Mendeley Data, DOI 10.17632/9tb7k297ff.1 https://doi.org/10.17632/9tb7k297ff.1 HealthyCropping com.keylorpineda.healthycropping is strictly non-commercial. It is distributed free of charge as an installable Android .apk exclusively through GitHub Releases https://github.com/keylorpineda/healthycropping/releases , with no advertisements or paid subscriptions. @software{pineda2026healthycropping, author = {Pineda, Keylor}, title = {{HealthyCropping: Offline-First Crop Disease Detection for Coffee, Cacao, and Banana}}, year = {2026}, url = {https://github.com/keylorpineda/healthycropping}, version = {0.1.0}, license = {MIT} } - Keylor Pineda — @keylorpineda https://github.com/keylorpineda - Jefferson Abarca — @JAbarca74 https://github.com/JAbarca74