Use Kitae as Your AI Coaching Companion

Kitae is built so its records read cleanly in any LLM. Combined with the project / custom GPT features in Claude and ChatGPT, you can turn it into a personal coach that learns your goals, watches your numbers over time, and gives feedback after every session.

Why this works

Project features (Claude Projects, ChatGPT Projects / Custom GPTs) keep your instructions and prior conversations across sessions. Set up your goals and coaching style once, then paste your daily Kitae log into the chat — the assistant already knows who you are, what you’re training for, and what you did last week.

One-time setup

1. Create a new project in Claude or ChatGPT.
2. In the project instructions, write:
   • Your training goal (hypertrophy, strength, fat loss, sport-specific, etc.)
   • Your training history and any injury notes
   • The exercises and equipment you have access to
   • The coaching style you want (strict and analytical, encouraging, periodization-focused, …)
3. Optionally paste a snapshot of your current PRs / 1RMs so the assistant has a baseline.

Daily workflow

1. Finish your session in Kitae.
2. Open the workout detail screen and tap the copy button (the doc.on.doc icon in the top-right toolbar). The icon turns into a checkmark to confirm.
3. Paste into your project chat and ask whatever you need — “How did today look?”, “What should I do next session?”, “Am I progressing on bench?”

Because the export is plain Markdown with exercise names, sets, RPE, estimated 1RM (Epley) and total volume, the assistant can reason about it directly without any extra parsing.

Include body composition (Apple Health)

In Settings → Apple Health, turn on Include body composition in copy and grant read permission. From then on, every copy automatically appends the latest body mass, body fat %, lean body mass, and BMI from Apple Health — each with the time it was measured, plus a 5-day moving average so day-to-day noise doesn’t mislead the analysis.

This lets you ask questions that connect training load and body composition — for example, “I’ve been adding volume but body fat is creeping up too. Should I switch to a cut?” — without having to type the numbers yourself.

If Apple Health isn’t authorized or a metric isn’t available, Kitae quietly skips it and copies the training portion as usual.

Example output

The export is plain Markdown — exercise names, sets, RPE, estimated 1RM (Epley), and total volume — so any modern LLM can read it directly, no extra parsing needed. (Kitae copies the section labels in Japanese, but you never need to translate anything before pasting.)

Training only (Apple Health toggle off):

## 2026-05-09 Workout Log

### Bench Press
- W 1: 60kg × 5 reps (RPE 5)
- Set 2: 80kg × 5 reps (RPE 7)
- Set 3: 100kg × 3 reps (RPE 9)
Estimated 1RM: 110kg
Total Volume: 1140kg

Training + body composition (Apple Health toggle on):

## 2026-05-09 Workout Log

### Bench Press
- W 1: 60kg × 5 reps (RPE 5)
- Set 2: 80kg × 5 reps (RPE 7)
- Set 3: 100kg × 3 reps (RPE 9)
Estimated 1RM: 110kg
Total Volume: 1140kg

## Body Composition (Apple Health)

- Body mass: 75.25 kg (measured 2026-05-09 14:30)
- Body mass, 5-day average: 75.10 kg
- Body fat: 15.3 % (measured 2026-05-09 14:30)
- Body fat, 5-day average: 15.5 %
- Lean body mass: 63.75 kg (measured 2026-05-09 14:30)
- Lean body mass, 5-day average: 63.50 kg
- BMI: 22.4 (measured 2026-05-09 14:30)
- BMI, 5-day average: 22.5

Bring the AI’s plan back into Kitae

The loop closes both ways. Ask your coach to reply with the next session (or a whole week) in the same Markdown shape Kitae uses, and you can import it directly instead of tapping it in by hand:

## 2026-05-12

### Bench Press
- Set 1: 80kg × 5 reps
- Set 2: 85kg × 5 reps
- Set 3: 90kg × 5 reps

1. On the home screen, tap + to start a new workout, then choose Import plan from text.
2. Paste the Markdown (clipboard or the built-in sample), review the preview, and tap Import. Several days can be imported at once.

A couple of things to tell your coach so the import succeeds:

• Each day starts with a ## YYYY-MM-DD heading. Only the date is read — any text after it is ignored, so ## 2026-05-12 and ## 2026-05-12 Push Day both work.
• Exercise names must match your Exercise Library exactly. Names that don’t exist yet are flagged in the preview — add them under Exercises first, or ask the coach to use your existing names.
• W marks a warmup set and Set a working set; × or x both work, and RPE is optional. Summary lines such as estimated 1RM or total volume are ignored on import, so it’s also fine to paste an exported log straight back.

A note on privacy

Kitae itself has no server — your workout records and any Apple Health data it reads stay on your device. The moment you paste into a chat, however, that text is sent to your LLM provider (Anthropic, OpenAI, etc.) under their respective terms. Decide what you’re comfortable sharing — for example, you can leave the HealthKit toggle off on days when you only want to share training numbers.

---

Operated by Femto Technologies, Inc. Contact: [email protected]