Un servidor del Protocolo de Contexto de Modelos (MCP) que conecta Claude a tu cuenta de Strava. Pídele a Claude en lenguaje natural que analice tu entrenamiento, prediga tiempos de carrera, calcule cargas de entrenamiento o genere un plan de entrenamiento periodizado completo.
13 herramientas en 5 categorías: autenticación, actividades, análisis, predicción, planificación
Flujo OAuth2 con servidor de callback local automático: sin necesidad de copiar códigos manualmente
Ritmos de entrenamiento basados en VDOT (Fórmula de carrera de Jack Daniels)
Métricas de carga de entrenamiento CTL/ATL/TSB (Carga de entrenamiento crónica/aguda, Equilibrio de estrés de entrenamiento)
Predicciones de tiempo de carrera mediante la fórmula de Riegel
Planes de entrenamiento periodizados completos (Base → Construcción → Pico → Taper) con lógica específica para la carrera
Related MCP server: Strava MCP Server
Node.js ≥ 18
Una cuenta de Strava
Claude Desktop (o cualquier cliente compatible con MCP)
Ve a strava.com/settings/api y crea una aplicación.
Dominio de callback de autorización: localhost
Anota tu Client ID y Client Secret.
cp .env.example .envEdita .env:
STRAVA_CLIENT_ID=your_client_id
STRAVA_CLIENT_SECRET=your_client_secret
STRAVA_REDIRECT_URI=http://localhost:8080/callback
TOKENS_FILE_PATH=./tokens.jsonnpm install
npm run buildEdita %APPDATA%\Claude\claude_desktop_config.json (Windows) o ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):
{
"mcpServers": {
"strava": {
"command": "node",
"args": ["C:/path/to/McpStrava/dist/index.js"],
"env": {
"STRAVA_CLIENT_ID": "your_client_id",
"STRAVA_CLIENT_SECRET": "your_client_secret",
"TOKENS_FILE_PATH": "C:/path/to/McpStrava/tokens.json"
}
}
}
}Puedes usar tanto
.envcomo el bloqueenven la configuración de Claude Desktop: ambos funcionan.
Reinicia Claude Desktop y, en una conversación:
Llama a strava_get_auth_url: Claude devolverá una URL
Abre la URL en tu navegador y autoriza la aplicación en Strava
La página mostrará "✓ Authentification réussie !" y los tokens se guardarán automáticamente
Herramienta | Descripción |
| Genera la URL de OAuth2 e inicia el servidor de callback local |
| Alternativa manual: intercambia un código de autorización por tokens |
| Comprueba si los tokens son válidos y cuándo caducan |
Herramienta | Descripción |
| Lista actividades recientes (Carrera, Ciclismo, Caminata, Todas) con distancia, ritmo, FC |
| Detalle completo de una actividad: parciales por km, vueltas, calorías |
| Estadísticas globales de Strava: esta semana, este año, histórico |
Herramienta | Descripción |
| Desglose del volumen semanal + puntuación de consistencia |
| CTL (forma física) / ATL (fatiga) / TSB (frescura) mediante TRIMP |
| Distribución en 6 zonas de ritmo, verificación de la regla 80/20 |
Herramienta | Descripción |
| Predice tiempos de llegada mediante la fórmula de Riegel a partir de un esfuerzo de referencia |
| Calcula la puntuación VDOT + 5 zonas de ritmo de entrenamiento a partir de cualquier rendimiento en carrera |
Herramienta | Descripción |
| Plan periodizado completo desde hoy hasta el día de la carrera (Base/Construcción/Pico/Taper) |
| Genera solo las sesiones de la próxima semana para una fase determinada |
Fase | Enfoque | Intensidad |
Base | Base aeróbica | Carreras suaves, carrera larga, progresiones |
Construcción | Umbral de lactato | Tempo, carrera media suave, carrera larga |
Pico | VO2max + específico de carrera | Intervalos, tempo, carrera larga |
Taper | Frescura | Volumen reducido, calidad de mantenimiento |
Volumen inicial — combinación de tu promedio de 4 semanas en km y la estimación semanal en km derivada del CTL (robusto ante pausas por lesiones)
Volumen pico — 1.4× el actual, con mínimos específicos por carrera (5K: 40km, 10K: 50km, Media: 60km, Maratón: 80km)
Progresión semanal — limitada a +10% por semana para prevenir lesiones
Semanas de recuperación — automáticas cada 4 semanas dentro de cada fase (volumen × 0.8)
Maratón (Construcción y Pico): Las carreras largas incluyen una sección a ritmo de maratón (~45% de la carrera a ritmo de competición)
Profundidad del Taper: Los 5K reducen hasta el 80% del volumen pico, el Maratón hasta el 40% — las carreras más cortas necesitan menos recuperación
Intervalos: 6 × 1000m para 5K/10K, 5 × 1000m para Media/Maratón
Basado en la Fórmula de carrera de Jack Daniels. Zonas calculadas como una fracción del VDOT:
Zona | % de VDOT | Uso |
Suave | 65% | Carreras diarias, carrera larga |
Maratón | 80% | Secciones a ritmo de maratón |
Umbral | 86% | Carreras tempo |
Intervalo | 98% | Intervalos de VO2max |
Repetición | 105% | Trabajo de velocidad / progresiones |
strava_generate_training_plan admite un modo manual que evita Strava por completo. Pasa current_weekly_km y goal_time juntos y no se necesita cuenta de Strava; útil para generar planes para amigos desde tu propio Claude Desktop.
Parámetro | Descripción | Ejemplo |
| Distancia de carrera |
|
| Fecha de carrera (AAAA-MM-DD) |
|
| Tiempo objetivo |
|
| Kilometraje semanal actual |
|
Génère un plan marathon pour mon ami, il court 55km par semaine et vise 3h45, la course est le 18 octobre 2026Mon amie veut courir un semi-marathon en 1h50 le 2026-09-14, elle fait environ 40km par semaineCuando se proporcionan tanto current_weekly_km como goal_time, la herramienta omite todas las llamadas a la API de Strava. El resume en la respuesta incluirá source_volume: "Fourni manuellement" para confirmar qué modo se utilizó.
Modo Strava | Modo Manual | |
Se necesita cuenta de Strava | Sí | No |
Calibración de volumen | Promedio de 4 semanas + CTL | Valor que proporcionas |
Estimación de VDOT | De actividades recientes o | De |
Caso de uso | Tu propio entrenamiento | Amigos / atletas sin Strava |
# Watch mode (no build step needed)
npm run dev
# Build TypeScript
npm run build
# Run built server
npm start
# Clean build artifacts
npm run cleansrc/
├── index.ts # MCP server entry point
├── config.ts # Env vars, Strava constants, race distances
├── types.ts # Shared TypeScript interfaces
├── auth/
│ ├── oauth.ts # OAuth2 URL builder, token exchange
│ ├── tokenStore.ts # Load/save tokens.json, expiry check
│ ├── callbackServer.ts # Local HTTP server for OAuth redirect
│ └── authTools.ts # MCP auth tools
├── strava/
│ ├── client.ts # Axios instance with auto token refresh
│ ├── activities.ts # Strava activities API
│ ├── athlete.ts # Strava athlete/stats API
│ └── activityTools.ts # MCP activity tools
├── analytics/
│ ├── metrics.ts # Weekly stats, pace zones, consistency score
│ ├── trainingLoad.ts # TRIMP, CTL/ATL/TSB computation
│ └── analysisTools.ts # MCP analysis tools
├── prediction/
│ ├── riegel.ts # Riegel race time prediction formula
│ ├── vdot.ts # VDOT computation, training paces, race equivalents
│ └── predictionTools.ts # MCP prediction tools
└── planning/
├── workouts.ts # Workout templates and distance bounds
├── plan.ts # Plan generation, phase allocation, VDOT estimation
└── planTools.ts # MCP planning toolstokens.json almacena tus tokens de acceso y actualización de Strava. Está en .gitignore: nunca lo subas al repositorio. Los tokens se actualizan automáticamente cuando caducan (los tokens de acceso de Strava duran 6 horas).
Modo Strava (tu propia cuenta)
Analyse mes 8 dernières semaines d'entraînementQuelle serait mon heure sur un marathon si je cours un 10K en 45min ?Génère-moi un plan d'entraînement pour un semi-marathon le 2026-09-20Calcule ma charge d'entraînement actuelle et dis-moi si je suis en forme pour une course ce week-endMontre-moi la répartition de mes allures sur les 4 dernières semainesModo manual (amigos / sin Strava)
Génère un plan marathon pour mon ami, il court 55km par semaine et vise 3h45, la course est le 18 octobre 2026Mon amie veut courir un semi-marathon en 1h50 le 14 septembre 2026, elle fait 40km par semaineMIT
Analyse la charge d'entraînement récente : volume hebdomadaire (km, temps, sorties), longue sortie, allure moyenne par semaine, score de régularité.
| Name | Required | Description | Default |
|---|---|---|---|
| weeks | No | Nombre de semaines récentes à analyser |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided; description does not disclose side effects, auth needs, or whether it's read-only, leaving behavioral transparency insufficient.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
One concise sentence with all necessary information, no wasted words.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a single-parameter tool with no output schema, the description adequately lists output metrics; missing return format is acceptable given low complexity.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% with a clear parameter description; the tool description adds no extra meaning beyond the schema, meeting baseline.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool analyzes recent training load with specific metrics (volume, long run, average pace, regularity), distinguishing it from siblings.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The context is implied but no explicit guidance on when to use this tool versus alternatives like strava_training_load or strava_weekly_workout.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Récupère les statistiques globales Strava : kilomètres cette semaine, ce mois, cette année, et all-time.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description reveals it is a read-only retrieval (km summaries) but does not mention authentication requirements, potential errors, or that it returns data only for the authenticated athlete. It is adequate but leans on the tool name for context.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
A single sentence with 15 words, front-loaded with the action verb 'Récupère' (retrieves). No wasted information.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
The description clearly lists the returned data (kilometers by week, month, year, all-time) which is complete for a simple stats tool. However, it could note that it returns stats for the authenticated athlete, but that is implicit in the tool name.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
No parameters exist, so schema coverage is 100%. The description does not need to add parameter details. Baseline 3 applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states it retrieves global Strava statistics for distances across four time periods. This distinguishes it from sibling tools like strava_get_activities which retrieves individual activities, and strava_training_load which focuses on training load.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies use for obtaining overall distance stats but does not explicitly contrast with alternatives or state when to prefer this tool. No when-not or prerequisites are provided.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Vérifie si des tokens Strava valides existent et quand ils expirent.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations exist, so the description bears full responsibility. It only mentions the check for token validity and expiry, lacking details on side effects, permissions, rate limits, or whether it is read-only.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
A single, clear sentence that efficiently conveys the purpose without wasted words.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
The description covers the core function but omits details about the output format (e.g., boolean or expiry timestamp). For a tool with no parameters, it is adequate but not fully explicit.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
With zero parameters and full schema coverage, the description adds value by explaining what the tool checks. The baseline for 0 params is 4, and the description meets it.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb ('Vérifie') and the resource (validité et expiration des tokens), distinguishing this utility tool from siblings that focus on activities, stats, or training plans.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No explicit when-to-use or when-not-to-use guidance is provided. The description implies usage for checking authentication status, but does not compare to alternative tools or state prerequisites.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Échange le code OAuth2 (depuis l'URL de redirection après autorisation dans le navigateur) contre des tokens d'accès et de rafraîchissement. Les tokens sont sauvegardés localement.
| Name | Required | Description | Default |
|---|---|---|---|
| code | Yes | Le code d'autorisation depuis l'URL de redirection Strava |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description carries full burden. It discloses side effect (tokens saved locally) but lacks details on idempotency, error handling, or overwrite behavior. Incomplete for a mutation tool.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences, no fluff, front-loaded. Could be slightly more structured but very concise.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Adequate for simple tool with one parameter and no output schema; explains process and side effect. Lacks error handling or prerequisite details, but overall sufficient.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema covers parameter `code` 100%, and description adds context (where code comes from: redirect URL after authorization), enhancing meaning beyond schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the action (exchange authorization code) and resource (tokens), specifying access and refresh tokens. It distinguishes from siblings like strava_get_auth_url which provides the code.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Implies usage after browser authorization but does not explicitly mention prerequisites (e.g., need code from strava_get_auth_url) or when not to use. No alternatives listed.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Génère un plan d'entraînement structuré depuis aujourd'hui jusqu'au jour de la course, en 4 phases : Base (fondation aérobie), Build (développement du seuil), Peak (VO2max et spécifique course), Taper (affûtage). Mode Strava : calibre automatiquement le volume et le VDOT depuis l'historique. Mode manuel : passer current_weekly_km + goal_time pour générer un plan sans compte Strava.
| Name | Required | Description | Default |
|---|---|---|---|
| target_race | Yes | Distance de la course cible | |
| race_date | Yes | Date de la course au format YYYY-MM-DD | |
| goal_time | No | Temps objectif au format 'H:MM:SS' ou 'M:SS'. Si absent, estimé depuis les activités récentes Strava. | |
| current_weekly_km | No | Volume hebdomadaire actuel en km. Si fourni avec goal_time, bypasse complètement Strava (mode manuel). | |
| runs_per_week | No | Nombre de sorties par semaine |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries full responsibility. It explains what the tool generates (plan in 4 phases) and the two input modes. However, it does not disclose authentication requirements, whether the tool modifies any data, or error handling for invalid race_date (e.g., past date). The description is adequate but not thorough.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is three sentences, each serving a distinct purpose: overall function and phases, Strava mode, manual mode. It is front-loaded with the core purpose and structured logically. No redundant or vague sentences.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given 5 parameters (2 required), no output schema, and no annotations, the description covers the key modes and parameter interactions. It could mention what the plan includes (e.g., daily workouts), but the phases and mode choice are well explained. The return format is not described, but the tool likely outputs a plan object, and the context is sufficient for an AI agent.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so baseline is 3. The description adds significant value by explaining the interaction between goal_time and current_weekly_km, including the condition for bypassing Strava (manual mode). It also clarifies default behavior for goal_time when absent. This goes beyond the individual parameter descriptions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool generates a structured training plan from today to race day in four phases. It distinguishes two modes (Strava auto-calibrate vs manual) and specifies the resource (training plan) and verb (génère). This differentiates it from sibling tools like strava_weekly_workout or strava_predict_race_time.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides explicit guidance on when to use each mode: Strava mode (auto-calibrate from history) and manual mode (requires current_weekly_km + goal_time). It does not explicitly state when not to use the tool or contrast with alternatives like strava_predict_race_time, but the context is clear enough for most use cases.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Récupère la liste des activités récentes de l'athlète avec distance, temps, allure et fréquence cardiaque.
| Name | Required | Description | Default |
|---|---|---|---|
| days | No | Nombre de jours passés à récupérer (défaut : 90) | |
| activity_type | No | Filtrer par type d'activité | All |
| max_results | No | Nombre maximum d'activités à retourner |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description must convey behavioral traits. It indicates a read operation (listing recent activities) with specific output fields, but does not mention authentication requirements, rate limits, or whether the data is user-specific. This is adequate but misses opportunities for richer disclosure.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
A single, well-structured sentence that front-loads the core purpose and expected output. No unnecessary words or repetition. Every element serves to inform the agent.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's low complexity (3 parameters, no output schema, no nested objects), the description provides a reasonable idea of what is returned (list of recent activities with key metrics). It could briefly mention pagination or limits, but it is mostly complete for its purpose.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100% with clear parameter descriptions for days, activity_type, and max_results. The description adds that the output includes distance, time, pace, and heart rate, but this does not enhance parameter semantics beyond what the schema already provides.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb 'Récupère' (retrieve), the resource 'liste des activités récentes de l'athlète', and includes specific metrics (distance, temps, allure, fréquence cardiaque). It distinguishes from siblings like strava_get_activity_detail (single activity) and strava_athlete_stats (stats).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance on when to use this tool versus alternatives such as strava_analyze_training or strava_weekly_workout. The description does not mention when not to use or suggest any prerequisites.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Récupère le détail complet d'une activité Strava (laps, splits kilomètre par kilomètre, calories).
| Name | Required | Description | Default |
|---|---|---|---|
| activity_id | Yes | ID de l'activité Strava |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries full burden for behavioral disclosure. It only states what the tool does, not side effects, authentication requirements, or limits. 'Complete detail' hints at large responses but is not explicit.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single concise sentence with no wasted words. It lacks structural separation (e.g., bullet points) but is efficient.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the simple schema and no output schema, the description covers the main purpose. However, it misses context like authentication prerequisites or that the activity must belong to the user, which affects completeness.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema covers the single parameter with a description (activity_id). The description adds no extra meaning beyond the schema, so baseline of 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool retrieves complete details of a Strava activity, listing specific data (laps, splits, calories). It distinguishes from sibling tools like strava_get_activities by focusing on a single activity's detailed data.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance on when to use this tool versus alternatives. It does not mention when not to use it or suggest siblings for different use cases, relying on implicit understanding.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Génère l'URL OAuth2 Strava. Ouvrir cette URL dans un navigateur, autoriser l'application, puis copier le paramètre 'code' depuis l'URL de redirection et le passer à strava_exchange_token.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations, but description conveys it generates a URL for manual browser interaction. It does not mention any side effects, but generating a URL is inherently safe. Could add that it does not modify any data.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences, front-loaded with the verb, minimal waste. Every sentence adds value: what it does and how to proceed.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a tool with no parameters and no output schema, the description provides the essential flow. Could mention that the URL includes client config (likely server-side), but adequate as is.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
No parameters, so base score 4. Description adds process context beyond schema (which is empty), but no parameter-level details needed.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Clear verb 'génère' and resource 'URL OAuth2 Strava'. Distinguishes itself from siblings by being the auth URL generator, and logically leads to strava_exchange_token.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Explicitly describes the sequence: open in browser, authorize, copy 'code', and pass to strava_exchange_token. Provides clear when-to-use and how-to-use guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Analyse la répartition des kilomètres par zone d'allure (Récupération, Facile, Aérobie, Seuil, VO2max, Anaérobie). Aide à vérifier la règle 80/20 (80% en zones basses).
| Name | Required | Description | Default |
|---|---|---|---|
| weeks | No | Nombre de semaines à analyser | |
| threshold_pace_min_km | No | Allure au seuil lactique au format 'M:SS' par km (ex: '4:30'). Si absent, estimée depuis les activités récentes. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries full burden for behavioral traits. It states what the tool analyzes but omits details like data source (Strava activities), required authentication, return format, or any side effects. This is insufficient for an AI agent to understand the tool's behavior.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two concise sentences with no unnecessary words. The first sentence front-loads the verb and resource, making it immediately understandable.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's simplicity (2 optional parameters, no output schema), the description is adequate for a basic analysis tool. However, it lacks information on prerequisites (e.g., authentication) and output format, which would improve completeness.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so baseline is 3. The description does not add any information about the parameters (e.g., how threshold_pace_min_km is used or the implications of default weeks). It relies entirely on the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb 'analyse' and the resource 'répartition des kilomètres par zone d'allure', listing the specific zones. It also mentions the 80/20 rule, providing additional context. This distinguishes it from sibling tools like strava_training_load or strava_weekly_workout.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explicitly ties usage to verifying the 80/20 rule, giving a clear use case. However, it does not mention when not to use this tool or compare it to alternatives, leaving some ambiguity.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Prédit les temps sur les distances standard (5K, 10K, Semi, Marathon) en utilisant la formule de Riegel (T2 = T1 × (D2/D1)^1.06) depuis une performance récente. Auto-détecte la meilleure performance récente sur Strava, ou accepte une entrée manuelle.
| Name | Required | Description | Default |
|---|---|---|---|
| known_distance_m | No | Distance de la performance connue en mètres (ex: 10000 pour 10K). Si absent, auto-détectée depuis les activités récentes. | |
| known_time | No | Temps de la performance connue au format 'H:MM:SS' ou 'M:SS'. Si absent, auto-détecté depuis les activités récentes. | |
| target_distances | No | Distances cibles à prédire | |
| days_lookback | No | Jours d'historique pour auto-détecter la meilleure performance |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so the description carries the burden. It discloses the use of the Riegel formula, auto-detection from recent activities, and manual input acceptance. It does not discuss authorization or rate limits, but as a read-only prediction tool, the behavior is adequately transparent.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise (two sentences) and front-loaded with the key formula and purpose. Every sentence adds value without redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no output schema, the description covers core functionality well but does not specify the return format (e.g., list of predicted times per distance). This is a minor gap for a prediction tool with 4 optional parameters.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so the baseline is 3. The description adds context (formula, auto-detect vs manual, target distances) but does not significantly extend beyond what the schema already provides.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states it predicts race times for standard distances using the Riegel formula, and distinguishes from siblings like strava_analyze_training or strava_vdot by focusing on time prediction from a performance.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explains when to use it (predicting race times from a recent performance) and mentions auto-detection or manual input, but does not explicitly state when not to use it or declare alternatives among sibling tools.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Calcule les métriques de forme, fatigue et fraîcheur via des moyennes exponentielles. CTL (fitness chronique, 42j) / ATL (fatigue aiguë, 7j) / TSB = CTL - ATL (fraîcheur). TSB > +10 = frais pour une course ; TSB < -20 = surcharge.
| Name | Required | Description | Default |
|---|---|---|---|
| days | No | Jours d'historique (min 42 pour que CTL soit significatif) | |
| hr_rest | No | Fréquence cardiaque au repos (bpm) | |
| hr_max | No | Fréquence cardiaque maximale (bpm) | |
| show_daily | No | Afficher les données jour par jour (sinon juste la synthèse) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description should carry the full burden of disclosing behavior. It explains the calculation method and interpretation, but does not clarify whether the tool makes external API calls or is purely computational given parameters. No side effects or authorization needs are mentioned.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences, front-loaded with the primary purpose, followed by concise interpretation thresholds. No superfluous words; every sentence adds value.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's simplicity and lack of output schema, the description covers the key metrics and their practical use. It could optionally describe the return format (e.g., numeric values for CTL, ATL, TSB) but is otherwise complete for an agent to understand the tool's output.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, and the description adds meaningful context beyond field names, such as explaining why 'days' must be at least 42 for CTL significance and providing real-world interpretation of TSB thresholds. This enhances understanding of parameter significance.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly defines the tool as calculating training load metrics (CTL/ATL/TSB) via exponential averages, specifying each metric's time window and interpretation. This uniquely distinguishes it from sibling tools like strava_analyze_training or strava_vdot, which focus on other aspects.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides interpretation thresholds (TSB > +10 fresh, TSB < -20 overload), implying usage for assessing training readiness. However, it does not explicitly state when to use this tool versus alternatives like strava_analyze_training, nor does it mention any prerequisites or exclusions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Calcule ton score VDOT (indice de capacité aérobie de Jack Daniels) depuis une performance récente. Retourne : score VDOT, 5 zones d'allure d'entraînement (Facile, Marathon, Seuil, Intervalle, Répétition), et les équivalents de temps sur toutes les distances standard.
| Name | Required | Description | Default |
|---|---|---|---|
| distance_m | Yes | Distance de la performance en mètres (ex: 10000 pour 10K) | |
| time | Yes | Temps de la performance au format 'H:MM:SS' ou 'M:SS' (ex: '45:30' pour un 10K en 45min30) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description must bear the burden. It correctly states it calculates and returns values, implying no side effects, but does not explicitly confirm read-only behavior or mention authorization or rate limits.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences, front-loaded with the main action and output, with no wasted words. Efficient and clear.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no output schema, the description sufficiently lists return components (score, zones, time equivalents). It is adequate for a calculation tool with two simple parameters, though could detail output format slightly more.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema has 100% coverage with clear descriptions and examples for both parameters. The description adds context about performance and output but does not significantly enhance parameter meaning beyond the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states it calculates VDOT score and training paces from a recent performance, listing specific outputs (score, 5 zones, time equivalents). This is distinct from sibling tools like strava_pace_zones or strava_predict_race_time.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies use with a recent performance but does not provide explicit guidance on when to use versus alternatives like strava_predict_race_time, nor does it mention when not to use the tool.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Génère uniquement les séances d'entraînement de la semaine prochaine, basées sur le VDOT et le volume actuel. Plus léger que le plan complet.
| Name | Required | Description | Default |
|---|---|---|---|
| vdot | No | Score VDOT (si absent, estimé depuis Strava) | |
| weekly_km | No | Volume hebdomadaire cible en km (si absent, calculé depuis Strava) | |
| phase | No | Phase d'entraînement actuelle | Build |
| target_race | No | Course cible (influence la structure des séances qualité) | 10K |
| runs_per_week | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, and the description does not disclose side effects, data access requirements, or whether the operation is destructive. It only mentions generation based on VDOT and volume, lacking behavioral details.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences, front-loading the main purpose in the first sentence. Every word is meaningful with no redundancy.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no output schema, the description should ideally describe the return format or content. It mentions generating sessions but omits what the output contains. With 5 parameters and 80% schema coverage, it is moderately complete but lacks return value information.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema covers 4 of 5 parameters with descriptions (80%). The tool description adds context by linking vdot and weekly_km to the overall purpose, and the phrase 'plus léger que le plan complet' hints at parameter scope, adding value beyond schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool generates next week's training sessions based on VDOT and volume, distinguishing it from the full plan sibling tool with 'plus léger que le plan complet'.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies when to use (for next week's sessions only, lighter than full plan), but does not explicitly state when not to use or provide alternative scenarios.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v1.0.0strava_analyze_trainingstrava_athlete_statsstrava_auth_statusstrava_exchange_tokenstrava_generate_training_planstrava_get_activitiesstrava_get_activity_detailstrava_get_auth_urlstrava_pace_zonesstrava_predict_race_timestrava_training_loadstrava_vdotstrava_weekly_workoutScored across 13 tools
Most tools target distinct aspects of training analysis (load, pace zones, VDOT, weekly plan), but some overlap exists between race time prediction tools (predict_race_time and vdot both estimate race times, though from different formulas). Descriptions help differentiate, but potential confusion remains.
All tools start with 'strava_' and use snake_case, but there is inconsistency in verb-noun order: some start with a verb (strava_analyze_training, strava_get_activities) and others with a noun (strava_athlete_stats, strava_pace_zones). This mixed pattern reduces predictability.
13 tools is a well-scoped set for a training analysis server. It covers authentication, activity retrieval, multiple analysis dimensions, and plan generation without being overwhelming.
The tool set provides comprehensive coverage of training analysis (load, pace, VDOT, race prediction, plan generation) and basic activity retrieval. Minor gaps exist, such as no tool to fetch or modify generated training plans, but core workflows are supported.
Connect Claude to your Intervals.icu watch data for fitness, workout review, and plan writing.
Turn Claude or ChatGPT into a cycling coach that plans your week, grades it, and adapts. Free beta.
Pace is a remote MCP server that exposes wearable and fitness data to Claude via the Model Context Protocol. It connects to Garmin, Oura, Whoop, Polar, Fitbit and 20+ devices and provides 15 tools for querying sleep, activity, recovery, and training data. Hosted on Google Cloud Run, OAuth 2.1 authentication, Streamable HTTP transport. Instructions: First you need to create an account at: https://pacetraining.co and connect your wearables. After that you can connect the remote Server via Custom Connector in Claude and OAuth 2.1 Flow startet.
Garmin data in Claude & ChatGPT via the Garmin Health API. OAuth sign-in, no password sharing.
