How to use Claude to log workouts through Telegram
IronClaude combines Claude Code, Telegram, Gemini transcription, and a private GitHub repository to log workouts and maintain context for interactive training plans.
news.ycombinator.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Log [exercise], [weight], [repetitions by set], and [effort rating] in my training history. Check for personal records and explain how this session relates to [current plan], [goals], and [recovery notes].
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- mosnicholas
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The author reported excellent structured workouts with progressive overload, deload weeks, and continuity between sessions.
- Human involvement
- Users provide goals, preferences, training logs, and weekly feedback; the author notes that the system cannot observe form like a real coach.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- The author built IronClaude around Claude Code with training data stored in a private GitHub repository.
- The user sends exercise details as a Telegram voice note for transcription.
- The system logs the workout, checks personal records, and responds with coaching context.
The author built IronClaude after becoming frustrated with repeatedly explaining their goals and preferences to workout chatbots. The application runs through Telegram, uses Claude Code as its agent, and keeps training information as Markdown and YAML files in a private GitHub repository owned by the user.
At the gym, the author sends a voice note describing an exercise, weight, repetitions, and effort. Gemini handles transcription. The system then logs the workout, checks for personal records, and responds with coaching context. The repository stores workout history, weekly plans, personal records, and the athlete profile, giving the agent persistent context across sessions.
Weekly planning is conversational. The bot asks about recovery, schedule changes, and training priorities before building a plan with progressive overload logic. The author says it can also account for recovery data when Whoop is connected. Setup includes collecting API keys, creating the private repository, an onboarding conversation, and deployment to Fly.io.
The author reported better structured workouts, including progressive overload, deload weeks, and continuity between sessions. They explicitly said it does not replace a coach who can observe exercise form.
Who it's for: This workflow may suit people who want conversational workout logging and persistent training records.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · mosnicholas
Frequently asked questions
Where does IronClaude store training history?
Workout history, plans, personal records, and profile information live as Markdown and YAML in a private GitHub repository owned by the user.
How are voice workout logs processed?
The user sends a Telegram voice note. Gemini transcribes it, and the system logs the workout, checks for personal records, and returns coaching context.
More details
Remaining steps
- The bot asks weekly planning questions and builds a plan using the user's training context.
About this account
Claude operates the coaching logic, scheduling, integrations, and file-based memory; users supply workout logs and planning answers.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=46977672. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: mosnicholas on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
Claude Code · Telegram Bot API · GitHub API · Gemini · Fly.io
Workout logging · Telegram · Persistent memory
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.