MCP App Store
by KeenEthics
SetRep icon

SetRep MCP Server for Claude and ChatGPT

By SetRep, Inc.

Overview

SetRep is a strength-training logbook and program planner for iPhone. The SetRep connector lets Claude read your SetRep account and, with SetRep Premium, change your training program when you ask. Claude can read your profile and units, your training locations and their equipment, your programs and their weekly schedule, your workout history, and your history on any exercise. It can search the SetRep exercise catalog and your custom exercises, and check which exercises you can do at a given location. Ask things like "How has my bench press moved over the last two months?", "What's my next workout?" or "Which exercises in my program can I do at my home gym?". Claude can swap, add, remove or reorder exercises, change sets and rep targets, and add, rename or reorder workouts. When Claude chooses any part of a change, such as the replacement for an exercise you named, it previews the change first: you see a line per change and how the weekly sets per muscle group move, and nothing is saved until you agree. A change you spell out completely is saved directly. Applying changes, creating custom exercises, adding exercise aliases and setting exercise notes require SetRep Premium. Reading your data and previewing changes are free. You need a SetRep account, created in the SetRep iOS app. Sign in on setrep.app with the same account to connect.

How to connect

  1. Claude: Open the SetRep listing in the Claude connectors directory and choose Connect. Claude listing SetRep exposes 14 tools here.
  2. ChatGPT: Open the SetRep listing in the ChatGPT app directory and choose Connect. ChatGPT listing SetRep exposes 14 tools here.

Tools

add_exercise_aliases

ChatGPT
Adds search aliases to one of the signed-in user's custom exercises and returns its search_aliases as stored with the added_aliases this call put there. Call it when the user says an exercise they already have goes by another name, such as a program from elsewhere listing it under a different one, so search_exercises finds it by that name next time. An alias the exercise already carries is skipped, and one that is another exercise's name or alias is refused with that exercise's id. Custom exercises only: a catalog exercise's aliases are maintained in the catalog, so use it as it is. Nothing else about the exercise changes; edits and deletes stay in the app.

apply_program_change

ChatGPT
Applies a change list to one of the signed-in user's programs in one transaction and returns the summary, warnings and volume changes preview_program_change gives; the app shows the change when it next comes to the foreground. Call it directly for an edit the user fully specified on one workout; for a list the user confirmed from its preview, pass the same list with its preview_id. It refuses a list that adds or removes a workout or touches more than one without a preview_id, an expired preview_id or one made for another list, and a preview whose program has changed since. Never apply a previewed list the user has not confirmed.

create_custom_exercise

ChatGPT
Creates a custom exercise for the signed-in user and returns it as search_exercises lists it, its new id ready for a program change; the app lists it after its next sync. Call it when the user wants to program or log an exercise and search_exercises finds no catalog or custom exercise for it under its name or a common alias; it refuses a name an existing exercise already carries. Fill the fields from what the user said and what the movement is: exercise_type decides what a set logs, primary_muscles drive the volume count and the search, equipment decides where the user can do it. Editing or deleting a custom exercise stays in the app.

fetch_exercise_availability

ChatGPT
Returns one exercise's availability at every training location of the signed-in user, default location first: per location its id, name, is_default, availability (available, inconvenient or unavailable), missing_equipment and training_location_exercise_override (the override that set it, explicit or inferred; null when the equipment rule decided). Call it when the user asks about one named exercise at a place: whether they can do it there, where they can, or why it is not suggested. Which exercises fit a location is search_exercises with training_location_id; a location's gear is fetch_locations.

fetch_exercise_history

ChatGPT
Returns the signed-in user's history on one exercise: its sessions newest first within since/until (default 12 weeks; has_more flags older ones), each with its date, location, the prescription and every set's targets, logged load, reps and rir. one_rep_max_series is the best Epley estimate per session, oldest first; personal_records the all-time bests, both in weight_unit. Call it when the user asks whether a lift is stuck or improving, how it has moved over a period, or what they did last time they did an exercise, asked by exercise rather than by workout. A question about a named workout or day is fetch_workout_record's. Widen since when sessions are missing.

fetch_locations

ChatGPT
Returns every training location of the signed-in user with its equipment, the weights each piece of equipment offers there (weight_profiles: an enumerated ladder or an increment), and per-exercise availability overrides. is_default marks the location the app proposes for the next workout. Call it before suggesting an exercise or a load so the suggestion matches the gear on hand.

fetch_profile

ChatGPT
Returns the signed-in user's age (whole years as of today's UTC date), gender, body weight, experience level and the units they log in; a field is null when the user has not set it. Call it before advising on loads, volume or exercise selection so the advice fits the user. It holds no training goal: a goal belongs to a program, so fetch_programs says what the user is training for. Body weight is one current value for sizing loads, not a basis for estimating calories, nutrition or recovery: when SetRep does not hold what a question needs (calories, nutrition, sleep, heart rate, a body-weight history), say so and do not estimate it from the data it does hold.

fetch_program

ChatGPT
Returns one program as its full tree: mesocycles, each with its workout_templates (exercise_templates with set_templates, exercise_preference and exercise_pinned_note inline) and its microcycles. Each mesocycle's volume holds weekly sets per muscle group (a secondary muscle counts half, warm-ups none); answer set-count and balance questions from it rather than counting. remaining_workouts lists this week's uncompleted slots: under ordered scheduling the first is next, under flexible the user picks any, never before the training day after last_completed_at. Call it when the user asks what to do next or what a program contains, and before any program change: every id a change takes comes from here.

fetch_programs

ChatGPT
Lists every program of the signed-in user with its goal and program_enrollment; is_enrolled marks the one in progress. Call it to find a program by name, to answer what the user is training for, or for the program_id that fetch_program and the program change tools take.

fetch_workout_record

ChatGPT
Returns one workout record in full: its exercise_records in order, each carrying the prescription the session was given (target_set_count and the target range) and its set_records with the targets each set was given (target_reps, target_weight_value, target_distance_value) next to what was logged (reps, weight_value, duration_seconds, distance_value, rir, rpe). A set whose completed_at is null was not done. Call it when the user asks about a specific workout, such as what they lifted on a given day; the workout_record_id comes from fetch_workout_records.

fetch_workout_records

ChatGPT
Lists the signed-in user's workout records newest first, each with its started_at, label, program, location, status (completed or inProgress) and the exercises it holds (exercise_records, without sets). since, until, program_id and exercise_id narrow the list; pages hold 50 and continue by passing next_cursor back as cursor. Call it when the user asks what they trained on a day or in a period, how often they trained, or when they last did an exercise, and to find the workout_record_id fetch_workout_record takes. For one exercise's sets across its sessions, fetch_exercise_history answers in one call.

preview_program_change

ChatGPT
Validates a change list against one of the signed-in user's programs without changing it and returns a summary line per change, warnings, the weekly sets per muscle group the list moves, and a preview_id apply_program_change takes with the same program_id and changes for one hour. Call it when you chose any part of the edit, such as the replacement for an exercise the user named or a goal turned into changes; show them the summary and apply only after they say yes. A list that adds or removes a workout or touches more than one must be previewed. A rejected list names the failing entry; fix it and call again.

search_exercises

ChatGPT
Returns the catalog exercises and the signed-in user's custom exercises matching every given filter, each with its id, exercise_type, movement_pattern, muscles, required equipment, affinity and pinned_note. With training_location_id each result also carries its availability there and the availability filter applies; without it no location is assumed. Call it to find exercise ids, to answer which exercises work a muscle group or fit a movement pattern, and to list what the user can do at a location. Whether, where or why the user can do one named exercise is fetch_exercise_availability's question once its id is found.

set_exercise_note

ChatGPT
Sets, revises or clears the signed-in user's pinned note on one exercise and returns the exercise with the note as stored and previous_text, what it replaced. The pinned note is the standing cue the app shows with the exercise in every workout (setup, form, equipment preference), so call it when the user asks to pin, note, remember or be reminded of something on an exercise, or to clear such a note; text null clears it. One note per exercise and new text replaces the whole note, so fold an addition into the pinned_note search_exercises or fetch_program already shows. Reading a note is those tools' job; a program's description is apply_program_change's.

add_exercise_aliases

Claude

apply_program_change

Claude

create_custom_exercise

Claude

fetch_exercise_availability

Claude

fetch_exercise_history

Claude

fetch_locations

Claude

fetch_profile

Claude

fetch_program

Claude

fetch_programs

Claude

fetch_workout_record

Claude

fetch_workout_records

Claude

preview_program_change

Claude

search_exercises

Claude

set_exercise_note

Claude

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/setrep"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/setrep)