API Docs – workout.chat

Connect to Claude.ai

workout.chat supports the Model Context Protocol (MCP), which means you can connect it directly to Claude.ai and chat with your workout data in natural language — log exercises, check your history, share workouts, and more.

Adding workout.chat as a connector in Claude.ai

Steps

1

Open Claude.ai → Customize → Connectors

In the left sidebar of Claude.ai, click Customize, then select Connectors.

2

Click "Add custom connector"

Give it a name like My Workout Chat and enter the MCP server URL below.

3

Enter the MCP server URL

Copy and paste the URL below into the Remote MCP server URL field.

https://workout.chat/mcp

4

Authorise access

Click Add and Claude will open a login page. Sign in to workout.chat and click Allow to grant access.

✓

Start chatting with your workouts

Claude can now log exercises, review your history, track weight, and generate share links — just by asking.

Example prompts to try: "Log today's back day — 3×10 pull-ups, 3×8 barbell rows at 135 lbs" · "How much did I bench last week?" · "Share my last workout" · "Log my weight as 185 lbs"

Authentication

The REST API uses session cookies for browser access. For external/programmatic access, use a Bearer token in the Authorization header.

MCP clients (e.g. Claude Desktop) authenticate via OAuth 2.0 PKCE. See the MCP OAuth Setup section.

Creating a token

POST to /mcp-tokens while logged in to generate a Bearer token.

curl -X POST https://workout.chat/mcp-tokens \
  -H "Content-Type: application/json" \
  -b "<session-cookie>" \
  -d '{"label": "my-script"}'

# Response
{"token": "abc123...", "label": "my-script"}

Using a token

curl https://workout.chat/api/workouts \
  -H "Authorization: Bearer <token>"

Base URL

https://workout.chat

All REST endpoints are prefixed with /api. Responses are JSON.

Workouts

GET /api/workouts List recent workouts

Query param	Type	Description
limitopt	integer	Max results, default 50

[
  {
    "id": 5,
    "user_id": 1,
    "performed_at": "2026-03-08T00:00:00",
    "title": "Back Day",
    "notes": null,
    "share_token": null,
    "created_at": "2026-03-08T20:57:33"
  }
]

POST /api/workouts Create a workout

Body field	Type	Description
performed_at	string	ISO datetime, e.g. `2026-03-08T09:00:00`
titleopt	string	Workout name
notesopt	string	Free-form notes
raw_textopt	string	Original unparsed log text

{"id": 14}

GET /api/workouts/{id} Get workout + exercises

PATCH /api/workouts/{id} Update title / notes / date

Updatable fields: performed_at, title, notes, raw_text

DELETE /api/workouts/{id} Delete a workout

Exercises

POST /api/workouts/{id}/exercises Add an exercise to a workout

Body field	Type	Description
name	string	Exercise name
setsopt	integer
repsopt	integer
entered_weightopt	number	Weight as entered by user
entered_weight_unitopt	string	`"lbs"` or `"kg"`
weight_kgopt	number	Normalised weight in kg
duration_secondsopt	integer
distance_kmopt	number
notesopt	string
positionopt	integer	Order within workout, default 0

{"id": 42}

DELETE /api/exercises/{id} Remove an exercise

Weigh-ins

GET /api/weighins List recent weigh-ins

Query param	Type	Description
limitopt	integer	Max results, default 30

POST /api/weighins Log a weigh-in (upserts on date)

Body field	Type	Description
date	string	ISO date, e.g. `2026-03-08`
weight_kg	number	Weight in kilograms

{"id": 7}

DELETE /api/weighins/{id} Delete a weigh-in

Profile

GET /api/profile Get current user's profile

{
  "id": 1,
  "name": "Sean",
  "email": "sean@example.com",
  "birth_year": 1990,
  "gender": "male",
  "height_cm": 178,
  "system": "imperial",
  "created_at": "2026-03-01T12:00:00"
}

PATCH /api/profile Update profile fields

Updatable fields: name, birth_year, gender, height_cm, system (metric or imperial)

{"ok": true}

MCP — Overview

workout.chat exposes a Model Context Protocol server at https://workout.chat/mcp. It accepts JSON-RPC 2.0 POST requests and is compatible with Claude Desktop, Cursor, and any other MCP client.

The MCP endpoint requires a valid Bearer token obtained via OAuth. See below.

Discovery

https://workout.chat/.well-known/oauth-authorization-server
https://workout.chat/.well-known/oauth-protected-resource

MCP — OAuth Setup

The server implements OAuth 2.0 Authorization Code flow with PKCE and Dynamic Client Registration (RFC 7591). MCP clients like Claude Desktop handle this automatically.

Endpoints

POST /oauth/register Dynamic Client Registration

curl -X POST https://workout.chat/oauth/register \
  -H "Content-Type: application/json" \
  -d '{"client_name": "My App", "redirect_uris": ["https://myapp.com/callback"]}'

# Response
{
  "client_id": "a1b2c3d4...",
  "client_name": "My App",
  "redirect_uris": ["https://myapp.com/callback"],
  "grant_types": ["authorization_code"],
  "token_endpoint_auth_method": "none"
}

GET /oauth/authorize Authorization endpoint (shows consent UI)

Query param	Description
client_id	From registration
redirect_uri	Must match registered URI
code_challenge	PKCE S256 challenge
code_challenge_method	`S256`
stateopt	CSRF token

POST /oauth/token Exchange auth code for access token

Body field	Description
code	Auth code from redirect
code_verifier	PKCE verifier
grant_type	`authorization_code`

{"access_token": "abc123...", "token_type": "Bearer"}

MCP — Tools

These tools are available to any connected AI assistant. All tools operate on the authenticated user's data only.

Workouts

log_workout write

Log exercises for a workout session. By default appends to the most recent workout from the same day (within 4 hours) to keep exercises grouped. Set new_workout=true to force a new session.

Parameter	Type	Description
exercises	array	List of exercise objects (see below)
performed_atopt	string	ISO date, defaults to today
titleopt	string	Workout name, e.g. "Push Day"
notesopt	string	Free-form notes
new_workoutopt	boolean	Force a new workout record

Each exercise object:

Field	Type	Description
name	string	Exercise name
setsopt	integer
repsopt	integer
weightopt	number
weight_unitopt	string	`"lbs"` or `"kg"`, default `"lbs"`
duration_secondsopt	integer
distance_kmopt	number
notesopt	string

list_workouts read

List the user's most recent workouts.

Parameter	Type	Description
limitopt	integer	Max results, default 10

get_workout read

Get a specific workout with all its exercises. Exercise IDs are included so they can be referenced for updates.

Parameter	Type	Description
workout_id	integer

update_exercise write

Correct a previously logged exercise. Only pass the fields that need to change — all others are left as-is.

Parameter	Type	Description
exercise_id	integer	From `log_workout` or `get_workout`
nameopt	string
setsopt	integer
repsopt	integer
weightopt	number
weight_unitopt	string
duration_secondsopt	integer
distance_kmopt	number
notesopt	string

remove_exercise write

Delete a mistakenly logged exercise entry.

Parameter	Type	Description
exercise_id	integer

share_workout write

Generate a public share link for a workout. Calling this multiple times for the same workout returns the same URL.

Parameter	Type	Description
workout_id	integer

{"url": "https://workout.chat/w/abc123xyz", "workout_id": 5}

Weigh-ins

log_weighin write

Record the user's body weight. Upserts on date — logging twice on the same day updates the existing entry.

Parameter	Type	Description
weight	number	Numeric value
unitopt	string	`"lbs"` or `"kg"`, default `"lbs"`
dateopt	string	ISO date, defaults to today

list_weighins read

List recent body weight entries.

Parameter	Type	Description
limitopt	integer	Max results, default 30

Profile

get_profile read

Get the current user's profile including name, height, birth year, and preferred unit system.

update_profile write

Update profile fields. Only pass fields that should change.

Parameter	Type	Description
nameopt	string
birth_yearopt	integer
genderopt	string	e.g. `"male"`, `"female"`
height_cmopt	integer
systemopt	string	`"metric"` or `"imperial"`

Utility

get_current_date read

Returns today's date and current time. Used to ensure workout dates are logged correctly.

{"date": "Monday, March 09, 2026", "time": "14:32"}