AI Upload Workouts

General
1

Feature Request: Muscular Load Input via Photo/Text Upload

Problem: Manual muscular load entry is tedious for complex workouts (CrossFit, functional fitness). This creates friction → lower compliance → incomplete data.

Proposed Solution: Accept workout input via photo (whiteboard) or text paste, then auto-map to WHOOP’s exercise library.

User Flow:

  1. Upload image or paste text (from SugarWOD, BTWB, coach notes)

  2. WHOOP parses and suggests matched exercises

  3. User confirms/edits → submits

Technical Approach (if helpful):

  • OCR (e.g., Google Vision, AWS Textract) for whiteboard photos

  • LLM or regex-based parsing for text normalization (“DB thruster” → “Dumbbell Thruster”)

  • Fuzzy matching against existing exercise taxonomy

  • Confidence threshold: auto-fill high-match items, flag ambiguous for user review

Value: Even 70% auto-match significantly improves UX for your most data-engaged users.

2

Hi @jdsteele , thanks for much for the feedback! I will pass it along to the team :slight_smile: