MCP App Store
by KeenEthics
Based Training icon

Based Training MCP Server for ChatGPT

By Based training

Overview

Based Training helps athletes review recorded activities and planned workouts, check training thresholds and connected-provider capabilities, get deterministic workout recommendations, create or reschedule structured workouts, and send compatible workouts to private connected fitness services when explicitly requested.

How to connect

  1. ChatGPT: Open the Based Training listing in the ChatGPT app directory and choose Connect. ChatGPT listing Based Training exposes 17 tools here.

Tools

create_from_library

ChatGPT
Use this when the athlete accepts a library workout or a get_recommendation result whose recommendation.kind is "library". Pass the library slug to add one personal workout to the Based Training plan. If the athlete has already subscribed their personal calendar to the Based Training feed, eligible scheduled workouts appear after that calendar refreshes. Do not call it for planned, rest, or conflict recommendations. Use the returned workout.id for any separately requested push_workout to an available training device or provider calendar.

create_workout

ChatGPT
Use this when the athlete explicitly asks to create a new custom workout rather than choose a library session. It validates a canonical nested step tree and creates one personal workout. Set scheduled_on and status "scheduled" to put it on the Based Training plan. If the athlete has already subscribed their personal calendar to the Based Training feed, eligible scheduled workouts appear after that calendar refreshes. Use the returned workout.id for any separately requested push_workout to an available training device or provider calendar.

detach_activity

ChatGPT
Use this only when the athlete confirms that a completed activity was matched to the wrong planned workout. It unlinks the activity and returns the workout to scheduled status; it does not delete either record.

get_activity

ChatGPT
Use this when an answer needs one completed activity's full recorded training metrics. Pass an activity ID returned by list_activities; this reads data only and does not sync providers.

get_recommendation

ChatGPT
Use this when the athlete asks what session to do next, including for a future date. It applies a deterministic safety policy to recent completed activities (granular types like ride count as cycling; run variants count as running) and the upcoming plan. Trust recommendation.recent_activities as the load the engine used; do not re-list history to second-guess it. Follow recommendation.kind: create_from_library only for "library"; reuse recommendation.workout.id for "planned" and reschedule only if the athlete accepts proposed_move; do not create anything for "rest" or "conflict" without an explicit choice. Call push_workout only when separately requested.

get_thresholds

ChatGPT
Use this when a training answer or workout prescription needs the athlete's effective FTP, running threshold pace, or threshold heart-rate anchors. Optionally resolve one sport or a past date.

get_workout

ChatGPT
Use this when an answer or edit needs one planned workout's full canonical step tree, scheduling details, session fueling, and provider push status. Pass an ID returned by list_workouts, get_recommendation, or a workout write tool. Fueling is derived from planned duration and peak intensity.

list_activities

ChatGPT
Use this to review the athlete's completed activity history over an optional date range or sport filter. Sport accepts a granular value (ride) or a family (cycling). Returns summary metrics and opaque IDs; use get_activity for one activity's full recorded metrics.

list_connections

ChatGPT
Use this before proposing a provider sync or workout push. Returns currently available connected provider names plus each provider's supported sports, metrics, and delivery constraints; it does not expose provider account identifiers or credentials.

list_library

ChatGPT
Use this when the athlete wants to browse reusable, validated Based Training workout templates, optionally by sport. This only reads the shared library; use create_from_library to add a selected template to the athlete's plan.

list_workouts

ChatGPT
Use this to inspect the athlete's planned workouts over a date range, optionally filtered by sport or status. Returns summaries and workout IDs; use get_workout before discussing or editing one workout's full step tree.

push_workout

ChatGPT
Use this only when the athlete explicitly asks to send an existing planned workout to an available connected training device or training-provider calendar within their private fitness-provider account. First call list_connections and choose a returned entry with capabilities.supported true, the workout sport in capabilities.sports, and every constraint satisfied. The destination is limited to the signed-in athlete's existing connected private account at the selected supported provider; it accepts no arbitrary recipient or URL and does not publish publicly. These tools cannot directly edit personal calendar apps. A workout within the next 90 days appears in a subscribed personal calendar only after that calendar refreshes the Based Training feed. Pass an existing workout.id; this queues provider-side workout creation that Based Training cannot undo.

reschedule

ChatGPT
Use this when the athlete asks to move an existing scheduled workout to a different date. It changes that workout's scheduled date on their private Based Training plan and queues updated delivery only to their available connected private fitness-provider accounts. It accepts no arbitrary external destination and does not publish publicly. If the athlete has already subscribed their personal calendar to the Based Training feed, the eligible event moves after that calendar refreshes.

set_threshold

ChatGPT
Use this when the athlete explicitly asks to record a new FTP, threshold heart-rate, or running threshold-pace anchor effective on a date. It appends a dated threshold record; it does not infer a value from activities.

sync_activities

ChatGPT
Use this when the athlete explicitly asks to refresh completed activities from a currently available connected pull-capable provider. It starts background provider requests for Strava, Polar, Karoo, Wahoo, or Garmin; poll list_activities for results. Wahoo also arrives by webhook. Apple Health imports on the iPhone when the app is open and when a workout is saved, and is not started here.

update_workout

ChatGPT
Use this when the athlete asks to edit an existing planned workout's steps, name, date, or status. It validates a canonical nested step tree. Personal and planned library workouts are overwritten in place; recommended auto-picks are copied into a new personal workout so the day's curated pick survives. Pass the returned workout.id to any separately requested push_workout.

whoami

ChatGPT
Use this to confirm which Based Training account is connected before reading or changing training data. Returns the athlete's email and connected provider names; it does not return internal identifiers.

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/based-training"><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/based-training)