MechAI

MechAI transforms customer calls into fully diagnosed repair jobs with AI-powered analysis and automated parts sourcing, so your mechanics can focus on fixing cars, not paperwork and collecting info.

Comment 1

Inspiration: Every auto repair shop faces the same frustrating bottleneck: phone calls. Mechanics spend hours on the phone taking down car details, describing symptoms, and trying to understand what's wrong—time they should be spending under the hood. We watched small shop owners juggle wrenches and phones, saw customers waiting days for callbacks, and knew there had to be a better way. MechAI was born from a simple question: What if AI could handle the entire customer intake process while providing mechanics with expert-level diagnostics?

What it does: MechAI is an AI-powered automotive diagnostic assistant that transforms how repair shops handle customer calls and job management.

When a customer calls:

AI Voice Agent answers (via Vapi/Telnyx) and conducts a natural conversation about their car problem Real-time transcription captures every detail as the customer speaks AI Information Extraction automatically pulls out: customer name, vehicle details (year/make/model), and symptoms Expert AI Diagnosis generates comprehensive analysis including: Root cause analysis with confidence levels Make/model-specific known issues and TSBs Step-by-step diagnostic procedures Required parts with cost estimates Safety warnings and alternative diagnoses Estimated repair time Automated Parts Sourcing uses AGI's browser automation to search suppliers and find parts Smart Dispatch matches jobs with nearby mechanics based on availability and distance Real-time Dashboard updates live via WebSockets so the shop sees jobs appear instantly The entire workflow—from call to diagnosis to parts search—happens automatically in under 30 seconds.

How we built it: We used Lovable for front end. Teelnyx as the voice agent and AGI for API agents.

Challenges we ran into: Comprehensive Diagnostics: Crafting AI prompts that generate truly useful, mechanic-grade diagnostics took many iterations. We added make/model-specific analysis, alternative diagnoses with probability ratings, and safety warnings.

Accomplishments that we're proud of: Zero Human Intervention: From call pickup to diagnosis to parts search—completely automated

✨ Sub-30-Second Diagnosis: AI generates expert-level analysis faster than any human technician ✨ Make/Model Intelligence: Diagnostics include vehicle-specific recalls, TSBs, and common issues ✨ Real-time Magic: Dashboard updates live as calls happen, feels like watching the future ✨ Multi-AI Integration: Combined Telnyx (voice), Lovable AI (diagnosis), and AGI (API) seamlessly ✨ Production-Ready: Built with edge functions, proper error handling, and scalable architecture

What we learned: Function calling is king: Getting structured output from LLMs using tool definitions is far more reliable than prompt engineering for JSON

Session-based APIs require different patterns: AGI's browser automation taught us about managing long-running async operations Real-time UX is transformative: Seeing data update live on the dashboard completely changes the user experience Edge functions scale: Serverless architecture means MechAI can handle 1 call or 10,000 without infrastructure changes

What's next for MechAI: SMS Photo Upload: Let customers text photos of their vehicle for visual diagnosis

🗺️ GPS Location Sharing: Customers share real-time location for mobile mechanic dispatch 📱 Mobile Mechanic App: Native app for mechanics to accept jobs, navigate, and update status 🔔 Smart Notifications: SMS/push alerts when jobs are assigned or status changes 📊 Analytics Dashboard: Track shop performance, average diagnosis time, revenue metrics 🔐 Customer Portal: Let customers track their job status and view diagnosis details

Built With

  • agi
  • lovable
  • telnyx
Share this project:

Updates