A growing set of small, focused AI tools for real life — count calories without spreadsheets, plan workouts without a coach, compare ETFs without a Bloomberg terminal. Same engineering rigor as the AI Lab. Built for everyone, not just enterprises.
Each app is a single focused tool — paste an input, get back something useful in seconds. Same architectural pattern as the AI Lab demos, just wrapped in a friendlier interface and pointed at a personal problem instead of an enterprise one.
Type what you did — "2h on the SPA schedules, 20 min call, chased the bundle" — and get properly drafted narratives with UTBMS codes, rounded to your firm's increment. The maths never touches the model, the codes are validated against the real UTBMS set, and the wording that gets written off is flagged before the bill goes out.
LiveLog a meal in plain text — "two eggs, toast, coffee" — and get a macro breakdown, vitamin estimate, and weekly trend. The food log that fills itself.
LiveA full training app in the browser. Build routines, then log every set live — rest timer, last session's numbers beside each set, plate maths, automatic PRs and 1RM. Plus an AI coach that writes your program, explains any lift, and reviews your week.
LivePaste 2–4 ETF tickers. Get a side-by-side comparison: fees, exposure overlap, correlation, dividend trail — and a recommendation tailored to your time horizon and risk tolerance.
LivePaste a set of property listings — copy-paste from Idealista, Rightmove, Zillow, or your notes. Each one gets classified deal / fair / overpriced / speculative / flag, with strengths, concerns, due diligence, a fit score, and a primary pick.
LiveThe interactive AI wiki for laundry. Three modes: Wash (garment + fabric → wash plan), Iron (heat, steam, technique), Stain Remover (any stain, any fabric). Step-by-step plans, do/don't lists, fabric symbols cheat sheet.
LiveWhat does an AI workload actually cost — in dollars AND in CO₂, water, and grid energy? Plug in tokens × model class × region, get back cost, kWh, CO₂e, litres, plus comparisons to cars, flights, households, and trees. Works for "is my prototype's bill reasonable" up to "what if someone burned 4 quintillion tokens?"
LiveTell it what's in your fridge. Get a recipe that actually works — with prep time, technique tips, and a shopping list for what's missing.
Log a week of sleep and lifestyle (caffeine, exercise, stress). One call surfaces patterns, predicts your best sleep window, and recommends the smallest change with the biggest impact.
Input destination, dates, planned activities. Get a smart packing list — climate-aware, activity-aware, and aware that you forget chargers.
The AI Lab shows what AI can do for enterprises: meeting intelligence, regulatory impact analysis, vendor scoring, strategic option weighing. Useful — but most of life isn't an enterprise.
AI Apps is the other half of the same idea: the same model, the same engineering rigor, applied to small, real, daily problems. You shouldn't need a finance degree to compare two ETFs, a nutritionist to understand a meal, or a coach to plan a sensible workout. The model already knows enough; we just need to put a clean interface in front of it and answer one specific question well.
Every app here will be free, run on the same single open-weight model, and ship with its internals visible — same transparency promise as the Lab. Built in public.
Stack: FastAPI · Nginx · Oracle Cloud · NVIDIA NIM · Nemotron 3 Super