Auto Mechanic AI Assistant App Development

Voice Wrench is a mobile app for auto mechanics. A technician logs in, starts one active repair project (VIN barcode scan or manual entry), and uses voice commands to interact with an AI agent during the repair. The agent provides torque specs, procedures, and other data from the AutoData API, showing results live on screen while speaking responses through Bluetooth earbuds. Closed projects are archived to the cloud. Build a cross-platform mobile app (Flutter or React Native) for iOS and Android, publishable to App Store and Google Play. Core functions: Passwordless email login (magic link or one-time code). One active repair project per user, stored locally on device (SQLite or similar). Project creation: radio choice “Scan VIN Barcode” or “Manual Entry”. Scan: use device camera to scan VIN barcode (ZXing/ML Kit), auto-fill VIN. Manual: text inputs for make, model, year, submodel. Repair session screen: shows vehicle details, continuous mic listening (always-on after start), live voice conversation with AI agent. Wake word: default “Wrench” (or “Hey Wrench”); optional custom wake word set by user in settings (saved per user). Pause/resume: tech says “[Name], hold on” to mute responses (still listens for wake word); “[Name], come back” to resume. Two-layer AI architecture: Voice layer: real-time mic input, wake-word detection, speech-to-text, TTS output, conversation flow, interruptions, Bluetooth earbud audio routing. Research layer: triggered by voice layer; makes secure OAuth 2.0 calls to AutoData API for torque specs, wiring diagrams, procedures, part verification; displays live search progress/results on screen (e.g. “Searching AutoData…” + results overlay); returns answer to voice layer for spoken response. On project close (marked completed), archive details to cloud/private server (Firebase Firestore or similar). All API calls require internet; no offline support. Security: encrypted local storage, HTTPS for network calls, OAuth 2.0 token management for AutoData, proper permissions for mic, camera (VIN scan), Bluetooth audio. Basic analytics/logging for pilot debugging. Provide quote for MVP development (pilot-ready, testable on devices). Include timeline, tech stack recommendation (Flutter vs React Native), and any assumptions/questions.

ID Projek: 40309788