get_training_overviewLABSCO SUMMARY
A coach that reads what you actually logged.
The read tools and the write tools point at the same record, so the programme responds to sessions you completed rather than sessions that were planned. Resolving exercise names to canonical ids before logging is what keeps a training history trendable instead of turning it into free text.
What it is
A hosted fitness coaching server at https://api.pelaris.io/mcp, with its tool set split between reads and writes. It holds your training profile, generates and manages programmes, plans and logs sessions, tracks benchmarks, body composition, goals and injuries, and returns coaching insight and a weekly debrief from what you have actually done.
What you get
- Training context, the active programme and recent sessions in one read (get_training_overview, get_active_program, get_session_details)
- A full programme generated from your profile and enrolled, with generation and programme status pollable (complete_intake, generate_program, get_generation_status, get_program_status, manage_program, get_onboarding_status)
- Sessions planned, modified, updated and deleted (create_planned_session, modify_training_session, update_session, delete_session, delete_sessions)
- Workouts logged as done, with exercises, sets and feedback (log_workout, log_completed_session), and exercise names resolved to canonical ids first (resolve_exercise_ids, swap_exercise)
- Benchmarks recorded and read back with history and trends (record_benchmark, get_benchmarks), alongside body composition and measurement trends (get_body_analysis)
- Goals listed and managed against linked benchmarks (list_goals, manage_goals)
- Coaching insight from your own training data, a generated weekly plan and a weekly debrief (get_coach_insight, generate_weekly_plan, get_weekly_debrief)
- Daily check-ins, injuries recorded, profile updates, feedback, and a search over curated training articles (daily_check_in, record_injury, update_profile, send_feedback, search_training_resources)
Requirements
An MCP client that can add a remote server: point it at https://api.pelaris.io/mcp. It supports OAuth with PKCE and Dynamic Client Registration; Claude's connector setup also wants the client id pelaris-claude under advanced settings.
Setup effort
Follow the published steps — ChatGPT: Settings → Apps → Add → enter the MCP Server URL https://api.pelaris.io/mcp
USE CASES
01
Getting a programme built around your situation
The intake captures your profile once, a full programme is generated and enrolled from it, and generation progress is pollable while it runs.
02
Logging a session as you finish it
Exercise names resolve to canonical ids first, then the session goes in with sets, targets and how it felt — in the same conversation, not in an app afterwards.
03
Adjusting a session that is too much today
Volume and intensity can be adjusted or an exercise swapped for an alternative, with the daily check-in giving the reason.
04
Finding out whether you are actually progressing
Benchmarks come back with history and trend, and body composition alongside them, so progress is read off numbers rather than off how the last session felt.
05
Reviewing the week and planning the next
The weekly debrief summarises what happened and sets the coaching focus, and the weekly plan is generated from it.
06
Training around an injury
An injury is recorded against your profile and goals are managed alongside it, so the programme is rebuilt with the limitation in it rather than around a gap in the log.
31 tools
Listed in project docs. Your client may show a different list.
get_active_programget_session_detailsresolve_exercise_idsget_benchmarksget_body_analysissearch_training_resourcesget_coach_insightget_onboarding_statusget_weekly_debriefget_generation_statusget_program_statuslist_goalscomplete_intakegenerate_programcreate_planned_sessionlog_workoutlog_completed_sessionswap_exercisemodify_training_sessionupdate_sessiondelete_sessiondelete_sessionsrecord_injuryupdate_profilesend_feedbackrecord_benchmarkdaily_check_inmanage_goalsmanage_programgenerate_weekly_planBEFORE YOU INSTALL
What to know first
Tools you get
31 tools named in the project's documentation.
INSTALLATION
Connect your client
Server setup
ChatGPT: Settings → Apps → Add → enter the MCP Server URL https://api.pelaris.io/mcp
Claude: Settings → Connectors → Add Custom → enter the MCP Server URL above → Advanced Settings → Client ID: pelaris-claude
Any MCP Client: connect to https://api.pelaris.io/mcp (OAuth 2.0 with PKCE and Dynamic Client Registration)
