AI Running Coach — personalized training via IM with automatic Strava-driven plan adjustments. Use when the user asks about running training, workout plans, marathon/half-marathon/10K/5K preparation, today's workout, weekly schedule, running stats/records (fastest, longest, weekly mileage), Strava analysis, or wants to chat with an AI running coach. Triggers on: running plan, training plan, today's workout, running stats, Strava data, marathon training, running coach.
OpenClaw-powered running coach that delivers personalized training via IM. The core value: Strava data flows in automatically, and the AI adjusts your plan in real-time.
The arc CLI handles all API calls. Token stored at ~/.config/airunningcoach/config.json.
If no token is configured (arc config show returns "not set"):
🏃 Welcome to AI Running Coach!
To connect your account:
1. Register → https://airunningcoach.net/register
2. Choose Pro (3-day free trial, cancel anytime)
3. Connect Strava → Profile page
4. Generate API Token → Profile page
5. Paste your arc_xxx token here
Already have an account? Just paste your token!
arc config set-token <token> then arc config testarc today # Today's workout
arc week # Full week view
Format the output in a friendly IM message. Add encouragement.
arc coach "user's message"
The coach endpoint has full context: active plan, Strava history, body feedback, personal records. Use this for ALL conversational interactions — it handles:
Do NOT ask for all parameters at once. Guide the user through a conversation:
Step 1 — Ask: "What distance are you training for? (5K / 10K / Half Marathon / Marathon)"
Step 2 — Ask: "How many weeks do you have? (4-16 weeks)"
Step 3 — Ask: "What's your goal? (Race a specific time / General fitness / Weight loss)"
Step 4 — Ask: "Any days you can't run? (e.g., Monday, Sunday)"
Step 5 — Confirm and generate:
arc plan create --race <type> --weeks <n> [--target <time>] [--goal <goal>] [--mileage <level>]
If user has Strava data, check their current mileage first:
arc strava summary
Then auto-fill --mileage based on their actual weekly average.
arc strava recent # Last 5 activities
arc strava summary # Full stats: records, PBs, weekly avg, monthly breakdown
Use strava summary when user asks about:
When user reports fatigue/injury:
arc feedback --type <fatigue|injury|soreness|illness> --severity <1-5> --message "description"
Then use arc coach to get adjusted recommendations. The coach considers all feedback.
When Strava data shows:
arc stats # Completion rate, streaks, totals
| Command | Purpose |
|---|---|
arc config set-token <t> | Save API token |
arc config test | Verify connection |
arc today | Today's workout |
arc week | This week's plan |
arc stats | Running statistics |
arc coach "msg" | Chat with AI coach |
arc plan create --race X --weeks N | Generate plan |
arc strava recent | Recent activities |
arc strava summary | Full Strava analytics |
arc feedback --type T --message "M" | Report body status |
arc config set-token <new_token> (regenerate at profile page)| Endpoint | Data Sent |
|---|---|
airunningcoach.net/api/v1/* | API token, training queries, plan parameters |
~/.config/airunningcoach/config.jsonThis skill sends running data and queries to airunningcoach.net. Install only if you trust this service. Privacy policy: https://airunningcoach.net/privacy