🏃‍♂️ Miles – Multi-Agent Running Assistant

Miles is an AI-powered multi-agent coach designed to help runners make smarter training decisions.
It combines Strava data, weather forecasts, and memory-aware reasoning to answer the big question every runner faces:

👉 "Should I run today?"

✨ Features

• Multi-agent orchestration using pydantic-graph

  • TeamLeader → Decides which sub-agent to activate
  • StravaCoach → Analyzes training load, pace, and recovery from Strava data
  • WeatherAssistant → Checks weather conditions before suggesting a run
  • RecoveryNode → Provides rest/recovery advice after hard efforts
  • TelegramFormatter → Outputs neatly formatted messages for chat delivery

• Memory persistence via Mem0

  • Stores and recalls per-user history (using Telegram chat_id)
  • Ensures context awareness across sessions

• Custom agent logic built with pydantic-ai

  • Structured outputs
  • Seamless integration with external APIs (Strava + Weather)

• Event-driven design ready for integration into Telegram bots or other chat interfaces.

---

Roadmap

•	Add GitHub Actions for daily automated run analysis
•	Predict next optimal run time with recovery modeling
•	Expand nutrition guidance from Strava performance data
•	Web dashboard with historical insights

🤝 Contributing

PRs are welcome! If you’d like to extend new agent nodes (nutrition, injury prevention, etc.), open an issue or draft a pull request.