MCP App Store
by KeenEthics
Kaizer icon

Kaizer MCP Server for ChatGPT

By Kaizer

Overview

Kaizer is the workspace personal trainers use to program and follow up their clients. Connect your Kaizer trainer account to work on your roster from ChatGPT. What you can do: - Find a client and review their profile, current routine, training status and workout history. - Generate a new routine with AI, continue from an existing one, or edit a routine in plain language; then send it to the client. - Adjust one exercise's sets, reps, load or RIR, and look up weight suggestions from the client's 1RM. - Follow progress: training analytics, per-exercise progression, weight history and physical test results. - Manage reminders, client notes and training plans (planifications). Who it is for: personal trainers and gym staff with a Kaizer trainer account. Clients (athletes) cannot connect. Limitations: every action works on data inside your Kaizer account only. AI routine generation runs in the background and can take a minute; you can ask for the result later. Deleting or overwriting data (routines, plans) asks for your confirmation first. Kaizer does not give medical diagnosis or treatment.

How to connect

  1. ChatGPT: Open the Kaizer listing in the ChatGPT app directory and choose Connect. ChatGPT listing Kaizer exposes 62 tools here.

Tools

addCalendarEventAttendee

ChatGPT
Book one of the trainer's clients onto an event. Idempotent — re-adding someone who cancelled confirms them again. Fails when the event is already at capacity.

addClient

ChatGPT
Add a client to the trainer's roster. With an email the client gets an invitation; without one they are created as a managed client with no account.

addClientNote

ChatGPT
Leave a note on a client. The client never sees it. Read notes back with getClientOverview.

archiveRoutine

ChatGPT
Archive a routine (works from any status), or restore an archived one to ready_for_client with archive: false. Returns { status }: archived, restored, already_archived, or not_archived. Archived routines only appear in getClientRoutines when status: "archived" is passed.

assignPlanificationToClient

ChatGPT
Assign a planification to a client (clones the base routine). Optional starts_at (YYYY-MM-DD).

cancelCalendarEventAttendee

ChatGPT
Take a client off an event. The row is kept as cancelled, so the client still shows up in getCalendarEventAttendees with that status.

createCalendarEvent

ChatGPT
Put an event on the trainer's calendar. starts_at/ends_at are full timestamps, not days. Book clients in with attendee_ids (they must be the trainer's clients — resolve names with findClientByName first). Pass repeat_frequency + repeat_until to make it recurring: one event per occurrence up to repeat_until (inclusive), capped at 366. The response is the FIRST occurrence; the rest share its group_id. is_enrollable lets clients sign themselves up — without it the event is private, visible only to its attendees; capacity caps confirmed attendees. Inside a gym the agenda is curated by its admins — a coach gets a 403 here (but keeps booking people in with addCalendarEventAttendee).

createExercise

ChatGPT
Create a new custom exercise in the trainer's exercise library (their pool of exercises). Use this when the trainer wants an exercise that isn't already in the catalog. Once created, it can be found via searchExercises and added to routines. Returns the new exercise's id and name. Inside a gym the library is the gym's and only its admins can add to it — a coach gets a 403, so don't offer.

createReminder

ChatGPT
Create a reminder that notifies the trainer at due_at. Use it when they ask to be reminded of something.

createTest

ChatGPT
Create a test in the trainer's library, not on one client. Check listTests first to avoid near-duplicates ("body fat" next to "% body fat").

deleteCalendarEvent

ChatGPT
Delete ONE event. Like updateCalendarEvent, a recurring event is not deleted as a whole — only the occurrence you name. Inside a gym only its admins can delete from the agenda; a coach gets a 403.

deletePlanification

ChatGPT
Delete a planification (already-assigned client routines are kept). Inside a gym any of its staff (admin or coach) can delete a program.

findClientByName

ChatGPT
Find a client_id from the roster (the trainer's clients, or the gym's members when working for a gym) by name, partial name, email, or tag. Returns { match }, { ambiguous }, or { not_found }.

findPlanificationByName

ChatGPT
Find a training_plan_id (planification) by name across the trainer's plans. Returns { match }, { ambiguous }, or { not_found }.

findRoutineByNameAndClientId

ChatGPT
Find a routine_id by name within one client. Returns { match }, { ambiguous }, or { not_found }.

getCalendarEventAttendees

ChatGPT
List who is on an event, oldest first. Cancelled attendees keep their row with status: "cancelled" so you can see who dropped out — count only the confirmed ones when reporting how full a class is.

getClientAnalytics

ChatGPT
Per-client training analytics over a date range (from/to, YYYY-MM-DD). Returns the metrics you list in metrics, or ALL of them when omitted — each with a plain-language description. Metrics: adherence, workouts_per_day, most_trained_exercises, total_volume, sets_per_muscle_group. For the whole roster use getTrainerMetrics; for one exercise's progression use getClientExerciseAnalytics.

getClientChatMessages

ChatGPT
The latest messages of the chat with a client, oldest first. sender is client for the client's messages and trainer for the ones sent from the trainer's side. A conversation_id of null means there is no chat with this client yet — reading never opens one.

getClientDetails

ChatGPT
Get client profile: name, email, birth_date, gender, goals, weak_points, weight, height, gym_experience, and tags (the trainer's own client-tag names).

getClientExerciseAnalytics

ChatGPT
One client's progression on ONE exercise over a date range (from/to). Metrics: estimated_weight (per-rep 1RM-style estimate — set rep, 1=1RM), volume and relative_strength (estimated weight ÷ current body weight) — one point per completed workout. Pass include_baseline: true to also get the last point before the range. Get exerciseId from getRoutineDetails or searchExercises.

getClientInstallments

ChatGPT
What one client still owes: their pending installments, oldest due first, each with its payments and reversals (a pending one has payments only when a reversal reopened it). Paid installments are not listed; past payments are in the history. Use it to get the installment_id recordClientPayment needs. next_installment_due_date is what recordClientPayment takes as pay_ahead_due_date; null while the charge is paused or has no price. Quote amount_due_formatted, never amount_due; paid_amount_formatted is what was already paid toward an installment in parts, which it owes that much less.

getClientOverview

ChatGPT
This-week snapshot for one client: adherence, days trained this week, the trainer's notes and attachment count. Use it for a quick 'how is X doing?' before diving into specific analytics.

getClientPayments

ChatGPT
Who owes the trainer money right now, one row per client. Quote the _formatted fields, never the raw amounts — those are in minor units, whose exponent depends on the currency. amount_formatted is the price of one period; what the client owes right now is owed_amount_formatted (0 while nothing is overdue). due_date is the OLDEST unpaid period's. Clients with payments disabled are absent entirely.

getClientRoutines

ChatGPT
List a client's routines (summary fields only). Archived routines are hidden unless status explicitly includes "archived".

getClientTests

ChatGPT
One client's measurements: every test they have values for, each with its series oldest-first. Tests with no values for this client are absent — listTests has those.

getClientTrainingStatus

ChatGPT
Get one client's name (first_name/last_name) and current routine_id, routine_status, last_trained_at, expires_at. Use when you already have the client_id.

getClientWeightHistory

ChatGPT
Get a client's dated weight measurements, newest first, in the trainer's own unit system. A scale reading also carries body_composition; getClientDetails already has the current weight.

getClientWorkoutHistory

ChatGPT
List workouts logged against a routine (any status).

getClientsTrainingStatus

ChatGPT
List the trainer's clients with their name (first_name/last_name) and current routine_id, routine_status, last_trained_at, expires_at. Use for 'who needs attention?' queries; for a single client use getClientTrainingStatus.

getMyProfile

ChatGPT
Get the current trainer's own profile (no args).

getOneRmPercentageWeight

ChatGPT
Turn a percentage of the client's 1RM on an exercise into a loadable weight. Use it when the trainer prescribes by percentage.

getPlanificationDetails

ChatGPT
Get a planification with its base routine and assigned clients.

getPlanifications

ChatGPT
Get all planifications created by the trainer. Archived ones are hidden unless status explicitly includes "archived".

getReminders

ChatGPT
The trainer's own reminders, soonest first. Filter by status, a due-date range, a text search, or the client they are about.

getRoutineDetails

ChatGPT
Get a routine's structure. light: true returns sessions per week + exercise/container names only (~10× smaller); default returns the full breakdown with sets/reps/weights (identical exercises across weeks collapse to one entry; weekly differences expand into a weekly array). Sessions come back as either session_groups (replicated across every week — edit with the grouped tools using session_group_id + session_node_group_id) or ungrouped_sessions (one-off, single-week — edit with the *ForSingleSession variants using session_id + session_node_id). Each session's content is a nodes tree where section/circuit/super_set containers carry children; even in light mode containers include their session_node_group_id and session_node_id, so use those to target a container as a parent (never invent an id). whiteboard nodes carry trainer notes in plain_text — surface them alongside exercises when describing a session so the trainer. An empty nodes array means no exercises yet. Sessions the client already trained are marked (trained_weeks on a group, trained: true on a session): they are the client's history and read-only — no edit, AI or manual, changes them.

getTrainerCalendar

ChatGPT
The trainer's calendar for a day range (from/to, YYYY-MM-DD): their events, which clients trained on which day, and the routines running out of programmed weeks in the range. Use include to ask for a single section when that's all you need. Events carry calendar_event_id — that's the id every other calendar tool takes. Ask for the bookings section to get who holds a spot in every session of the range at once, plus who waits in line for a full one; getCalendarEventAttendees answers for one event.

getTrainerClients

ChatGPT
List the trainer's clients with profile fields: name, email, gender, goals, weight, height, status, and tags (the trainer's own client-tag names; tag is the user handle). Same data as getClientDetails returns per client. Filter by tags with ids from listTags.

getTrainerMetrics

ChatGPT
Aggregate business/roster metrics for the signed-in trainer over a date range (from/to, YYYY-MM-DD; up to 730 days apart). Returns the metrics you list in metrics, or ALL of them when omitted — each with a plain-language description of what it measures. These cover the whole client roster, not a single client. Metrics: adherence, workouts_logged, total_clients, workouts_per_day, workouts_per_client, most_used_exercises, personal_records, time_saved.

getWeightSuggestion

ChatGPT
Suggest a starting weight for an exercise from the client's workout history. effort is read and answered in effort_scale (rir by default).

getWorkoutBySession

ChatGPT
Get a session's workout: exercises, sets, reps, weights.

listTags

ChatGPT
List the trainer's own tags (their custom labels for clients, exercises, templates and training plans). Call this before filtering by tags: searchExercises and getTrainerClients take tag ids, and this is the only place they come from. Each tag returns its id, name and type ("exercise", "training_plan_and_template" or "client") — only exercise tags work as an exercise filter and only client tags as a client filter.

listTests

ChatGPT
List the trainer's tests — the metrics they track on clients (e.g. "% body fat", "vertical jump"), with the unit each one is measured in. Test ids come from here.

logClientTestValue

ChatGPT
Record what a client measured on a given day. One value per test per day — a second one for the same day is rejected, and editing an existing value is dashboard-only.

recordClientPayment

ChatGPT
Record that a client paid. ONE call settles EXACTLY ONE period — the oldest unpaid one — so a client three periods behind needs three calls. Records are immutable once written.

requestNutritionPlanEdit

ChatGPT
Edit an existing nutrition plan in place with AI (same plan id). Describe the changes in free text; everything the instructions don't name stays as it is.

requestNutritionPlanGeneration

ChatGPT
Generate an AI nutrition plan for a client from their saved nutrition profile and targets. It lands as a draft for the trainer to review — it is never sent to the client.

requestPlanificationGeneration

ChatGPT
Generate an AI planification (template, not client-scoped). Status values: created = success; failed = tell the trainer; still_running = trainer can ask later. gym_experience: 1=beginner, 2=intermediate, 3=advanced. weeks is capped at 16 (the generator's limit) — for a longer program, generate a block of that length and tell the trainer to extend or regenerate it later. Inside a gym the programs are the gym's and only its admins can create one — a coach gets a 403, so don't offer.

requestRoutineEdit

ChatGPT
Edit an EXISTING routine with AI (DEFAULT for ANY change to a routine's content or structure: add/remove/swap exercises or sessions, circuits/supersets, reorder, rename, change progressions…). To continue a client's training in a new routine use requestRoutineFromExisting instead. Describe the changes in free text; everything the instructions don't name is preserved exactly — including each exercise's current weights. For a pure sets/reps/RIR tweak of ONE exercise prefer updateExercise. The routine is replaced in place (same routine_id), except the sessions the client already trained (trained_weeks / trained in getRoutineDetails): those stay exactly as they were. Status values: created = success (the updated routine is returned inline); failed = tell the trainer it failed; still_running = trainer can ask later.

requestRoutineFromExisting

ChatGPT
Generate a NEW AI routine for a client that builds on one of their existing routines: the AI starts from that routine's structure, exercises and loads. DEFAULT when the trainer wants to continue or progress a client's training. The source routine is left untouched (unlike requestRoutineEdit, which rewrites it in place). Take sessions_per_week, session_seconds and weeks from the source routine unless the trainer changes them; weeks is capped at 16.

requestRoutineGeneration

ChatGPT
Generate a complete AI routine for a client from scratch (DEFAULT — use this for generic 'create a routine' requests, unless the trainer explicitly asks for an empty/manual routine). To continue from a routine the client already has, use requestRoutineFromExisting. Status values: created = success; failed = tell the trainer it failed; still_running = trainer can ask later. Client profile fields are auto-filled.

searchExercises

ChatGPT
Search the exercise catalog by name, muscle group, equipment, difficulty, or trainer tags. Filters combine with AND across dimensions and match-any within a dimension. Use tags to search the trainer's own labels — it takes tag ids, so call listTags first to resolve a label's name into its id. Use difficulty (array of "beginner", "intermediate" and/or "advanced") to match any of those levels; exercises that have not been graded are excluded whenever it is set. Each result returns its name, muscle groups, equipment, difficulty and tags.

sendChatMessageToClient

ChatGPT
Send a text message to a client in the chat with them. The client gets it like any message typed from the app, push included; the chat is opened first if there is none yet.

sendPaymentReminder

ChatGPT
Send the client their payment reminder. The copy is a fixed Kaizer template in the client's language — it can't be worded by the trainer, so don't offer to draft it. Check can_remind on getClientPayments first; remind_blocked_reason says why it would fail.

sendRoutineToClient

ChatGPT
Deliver an already-generated routine to the client: transitions a trainer-only draft (status ready_for_trainer) to client-visible (ready_for_client). Use this when the trainer wants to send a routine that was generated as a draft (auto_send=false). Returns { status: "sent" } on success, { status: "already_sent" } if the client already has it, or { status: "not_sendable" } (with the current routine_status) if the routine isn't a sendable draft yet — e.g. still generating, failed, or canceled. Every outcome also carries routine_name, client_id and client_first_name/client_last_name for confirmation display.

setClientExerciseOneRm

ChatGPT
Record a 1RM the trainer states for a client on an exercise (a tested max, or one from before Kaizer). It becomes the standing 1RM getOneRmPercentageWeight prescribes from.

syncPlanification

ChatGPT
Re-sync the base routine to specific clients (replaces their routine sessions). Inside a gym any of its staff (admin or coach) can re-sync a program.

undoWorkout

ChatGPT
Undo a completed workout: the session goes back to untrained and its records are rolled back. Use it when a session was closed by mistake.

updateCalendarEvent

ChatGPT
Patch ONE event. Pass only the fields that change. A recurring event is not updated as a whole — this touches the single occurrence you name, not its group_id siblings. Inside a gym only its admins can reshape the agenda; a coach gets a 403.

updateClientProfile

ChatGPT
Update a client's profile with what the trainer states in chat ("pesa 96 kg"). Partial: send only the given fields. Read it back with getClientDetails.

updateExercise

ChatGPT
Update an exercise's sets/effort across every week. Pass the full set list. Skip effort on nodes whose show_effort is false. effort is in the unit of the exercise's effort_scale: to prescribe RPE 8 on an exercise stored in RIR, either send effort_scale: "rpe" with effort: 8 (the body is read in the scale it declares and the stored values are converted) or send effort: 2.

updateExerciseForSingleSession

ChatGPT
Update sets/effort in ONE session only (no propagation). Skip effort on nodes whose show_effort is false. effort is in the unit of the exercise's effort_scale; sending a different effort_scale converts the stored values and reads this body in the new scale. Exception: an exercise linked across weeks has one scale, so changing effort_scale also re-expresses every other week of that exercise in the new scale (same effort, new unit).

updatePlanification

ChatGPT
Rename, re-describe or archive a planification. status: "archived" archives the plan, its template routine and every client routine assigned from it; moving it back out of archived restores them. An archived plan can be neither assigned nor synced. Inside a gym the programs are the gym's and any of its staff (admin or coach) can edit them.

Capabilities

Writes

Example Prompts

Click any prompt to copy it.

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/kaizer"><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/kaizer)