AI Fitness Trainer
Ask anything about workouts, nutrition, body composition, recovery, or supplements. Powered by Qwen 2.5 running directly in your browser — no data is ever sent to a server.
Could not load the AI model. If your browser supports WebGPU or WebAssembly, try refreshing the page. Firefox, Chrome, and Edge are all supported.
Fitness Questions Answered
Click any question to see the answer, or ask the AI trainer below for a personalized response.
How much protein do I need per day?
▼
Most research recommends 1.6-2.2g of protein per kg of body weight for muscle building. For an 80kg person, that is 128-176g daily. Use our Protein Calculator for your exact needs and check your daily intake in the Habit Tracker.
How many calories should I eat to lose weight?
▼
A deficit of 300-500 calories below your Total Daily Energy Expenditure (TDEE) is recommended for sustainable fat loss. Calculate your TDEE with our TDEE Calculator, then use the Calorie Calculator to set your target intake.
What is a good workout routine for beginners?
▼
A full-body routine 3x per week is ideal for beginners. Focus on compound lifts: squats, deadlifts, bench press, rows, and overhead press. Start with 3 sets of 8-12 reps. See our Training Programs for structured plans and log your workouts with the Workout Log.
How long to see muscle growth?
▼
With consistent training and proper nutrition, most beginners see noticeable changes in 8-12 weeks. Track your transformation with Progress Photos and Measurement Log to see real data over time.
Initializing...
Your personal fitness AI is loading.
The model (~350MB) downloads once and caches for future visits.
Loading AI model
Downloading...
One-time download ~350MB. Cached locally for repeat visits.
Weight Loss
Workout Frequency
Pre & Post Workout
Hydration
Supplements
Muscle Balance
Your Tools
Privacy
100% private. Your questions are processed on this device. Nothing is sent to any server. Model weights are fetched from HuggingFace CDN and cached locally.
Model
Qwen2.5 0.5B by Alibaba. Runs via Transformers.js with WebGPU acceleration. Falls back to CPU on unsupported browsers.