AutoShop AI.
Shop management for mobile mechanics, built for techs by a tech — fourteen years of European automotive diagnostics turned into software that actually matches how the work happens.
Most shop software is built by people who've never had a customer standing over them while they diagnose a no-start in a parking lot. AutoShop AI is built from the other direction: I spent fourteen years as the tech, and every feature exists because the job needed it — not because a product manager imagined it.
The platform
It's true multi-tenant SaaS: self-serve onboarding spins up an isolated tenant, X-Tenant-ID isolation keeps every shop's data walled off, and the stack — FastAPI, PostgreSQL with pgvector, Redis and Celery — handles the async reality of a business that happens in driveways. Customers reach the shop through a hybrid IVR (neural TTS with automatic fallback) and a customer-facing PWA that works offline, because job sites don't come with good signal.
The detail techs actually love
A Manifest V3 browser extension that carries vehicle context across seven different parts-supplier sites — punch in the vehicle once, and every supplier search is pre-filled. It's a small thing that saves minutes on every single job, which is exactly the kind of feature you only think of if you've lived the workflow.
AI throughout follows the same rule as everything I ship: the AI drafts, extracts, and suggests — the tech keeps the judgment call.