Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/chrisdoc/hevy-mcp/llms.txt

Use this file to discover all available pages before exploring further.

Hevy MCP registers two server-side MCP prompts that coordinate common multi-step workflows. Unlike individual tools, prompts encapsulate a sequence of instructions that guide the assistant through a complete task — they are invoked by name from MCP clients that support prompts, and the server expands them into a structured message before the assistant begins.

​
analyze-workout-progress

Analyzes recent workout activity and body-measurement trends over a configurable number of weeks. When invoked, the prompt instructs the assistant to call the get-training-summary tool first, then produce a structured analysis grounded exclusively in the returned evidence. What the prompt does:
  1. Calls get-training-summary with the requested number of weeks to retrieve a compact, combined evidence set of recent workouts and body measurements.
  2. Instructs the assistant to use the returned period, workout frequency, volume, exercise variety, session list, and measurement trend fields — without issuing separate pagination or count calls.
  3. Guides the analysis across workout frequency, training volume, exercise variety, consistency, and body-measurement trends.
  4. Directs the assistant to distinguish observations from suggestions, note missing or limited data, and avoid unsupported claims or medical conclusions.
Arguments:
ArgumentTypeRequiredDefaultDescription
weeksinteger (1–12)No4Number of recent weeks to analyze.
Example invocation arguments: 
{
  "weeks": 6
}
With MCP SDK v1.29.0, clients invoking analyze-workout-progress with its default value must send arguments: {}. Omitting the entire arguments object is rejected by that SDK version before the default is applied.

​
create-workout-from-routine

Creates a completed workout entry in Hevy from an existing saved routine. The prompt loads the routine, maps plan fields to workout fields, then collects actual set results and end time from the user before calling create-workout — it will never invent completion data. What the prompt does:
  1. Calls get-routine with the provided routineId and maps supported plan fields: routine title to workout title, plus each exerciseTemplateId, supersetId, exercise notes, and set type.
  2. Explicitly skips routine-only fields (restSeconds, repRange) that are not valid in create-workout.
  3. Asks the user for actual completed set data for every set — including applicable weight, reps, distance, duration, RPE, or custom metric values — before proceeding.
  4. Collects the required endTime in strict UTC YYYY-MM-DDTHH:mm:ssZ format and confirms any other missing required fields.
  5. Once all data is confirmed, calls create-workout with only the fields that tool supports.
Arguments:
ArgumentTypeRequiredDescription
routineIdstringYesID of the saved routine to use as a guide.
startTimestring (UTC ISO 8601)YesWorkout start time in YYYY-MM-DDTHH:mm:ssZ format.
Example invocation arguments: 
{
  "routineId": "abc123",
  "startTime": "2025-01-15T09:00:00Z"
}
The prompt explicitly instructs the assistant never to invent completion data. If the actual set results or endTime are unavailable, the assistant will ask the user for them instead of creating the workout.

​
When to use prompts vs tools

Use prompts when you want a guided, multi-step workflow that coordinates several tool calls automatically. Prompts handle the sequencing logic for you — the assistant follows a structured set of instructions and knows what to ask or collect at each step. Use tools directly for simpler, single-step operations where you already know exactly what you want — for example, calling get-workouts to list recent sessions, get-routine to inspect a saved routine, or create-body-measurement to log a measurement. Individual tools are documented in the tools overview.

Tools Overview

Browse all 25 Hevy MCP tools organized by category.

Resources

Read workout count, exercise catalog, and profile data via MCP resource URIs.

Build docs developers (and LLMs) love

Get started for freeTalk to us