Skip to main content
Glama
llmkit-mcp-server

llmkit-mcp-server

by smigolsmigol

@f3d1/llmkit-mcp-server

Seguimiento de costes de IA para Claude Code, Cline, Cursor y Claude Desktop. 11 herramientas para consultas de gasto, presupuestos, costes de sesiones locales y atribución de agentes en 11 proveedores de LLM.

Parte de LLMKit, una puerta de enlace API de código abierto con seguimiento de costes y aplicación de presupuestos.

Inicio rápido

Añádelo a tu .mcp.json (Claude Code) o .cursor/mcp.json (Cursor):

{
  "mcpServers": {
    "llmkit": {
      "command": "npx",
      "args": ["@f3d1/llmkit-mcp-server"],
      "env": {
        "LLMKIT_API_KEY": "llmk_your_key_here"
      }
    }
  }
}

Las herramientas locales (llmkit_local_*) funcionan inmediatamente sin necesidad de una clave API. Detectan automáticamente los datos de Claude Code, Cline y Cursor en tu máquina. Para las herramientas de proxy, crea una clave gratuita en el panel de control.

Related MCP server: Langfuse MCP Server

Herramientas

Herramientas de proxy (requieren clave API)

Herramienta

Qué hace

llmkit_usage_stats

Gasto, solicitudes, modelos principales durante un periodo

llmkit_cost_query

Costes agrupados por proveedor, modelo, sesión o día

llmkit_budget_status

Límites presupuestarios y saldo restante

llmkit_session_summary

Sesiones recientes con coste, duración y modelos

llmkit_list_keys

Todas las claves con estado y fecha de creación

llmkit_health

Ping al proxy con tiempo de respuesta

Herramientas locales (no requieren clave)

Detectan automáticamente las herramientas de programación con IA instaladas y agregan datos de todas ellas.

Herramienta

Qué hace

llmkit_local_session

Coste de la sesión actual en todas las herramientas detectadas

llmkit_local_projects

Coste acumulado en todos los proyectos y sesiones

llmkit_local_cache

Análisis de ahorro por caché de prompts

llmkit_local_forecast

Proyección mensual frente a suscripción máxima

llmkit_local_agents

Atribución de costes de subagentes (Claude Code)

Hook SessionEnd

Registra automáticamente los costes de sesión al salir de Claude Code:

{
  "hooks": {
    "SessionEnd": [
      {
        "type": "command",
        "command": "npx @f3d1/llmkit-mcp-server --hook"
      }
    ]
  }
}

Variables de entorno

Variable

Requerida

Descripción

LLMKIT_API_KEY

No

Clave API para herramientas de proxy. Las herramientas locales funcionan sin ella.

LLMKIT_PROXY_URL

No

URL del proxy (por defecto al servicio alojado)

LLMKIT_CLINE_DIR

No

Sobrescribir la ruta del directorio de datos de Cline

Herramientas compatibles

Las herramientas locales detectan datos de:

  • Claude Code (~/.claude/projects/)

  • Cline (VS Code, Insiders, VSCodium, Cursor, Windsurf globalStorage)

  • Instalaciones WSL (escanea todas las distribuciones mediante rutas UNC en Windows)

  • VS Code Server / Cursor Server (extensiones remotas SSH/WSL)

Licencia

MIT

Available Tools

11 tools
llmkit_budget_statusA
Read-onlyIdempotent

Check budget limits and remaining balance

Inputs: budgetId. Returns: budgets

llmkit_cost_queryB
Read-onlyIdempotent

Query cost breakdown grouped by provider, model, session, or day

Inputs: groupBy (required), days, provider, model. Returns: groupBy, days, breakdown

llmkit_healthA
Read-onlyIdempotent

Check proxy health and response time

Inputs: no named parameters. Returns: status, responseTimeMs

llmkit_list_keysA
Read-onlyIdempotent

List all API keys with status and creation date

Inputs: no named parameters. Returns: keys

llmkit_local_agentsA
Read-onlyIdempotent

Subagent cost attribution for the current Claude Code session. Shows which agents cost the most.

Inputs: no named parameters. Returns: sessionId, totalCostUsd, mainConversationCostUsd, subagentsTotalCostUsd, +2 more

llmkit_local_cacheA
Read-onlyIdempotent

Cache savings analysis across all detected AI coding tools. Shows how much prompt caching saved.

Inputs: no named parameters. Returns: savings, totalSavedUsd

llmkit_local_forecastA
Read-onlyIdempotent

Monthly cost projection based on local AI tool usage. Compares to Max subscription.

Inputs: no named parameters. Returns: projectedMonthlyUsd, dailyAverageUsd, totalTrackedUsd, totalSessions, +1 more

llmkit_local_projectsA
Read-onlyIdempotent

Cumulative cost across all projects and sessions from all detected AI coding tools, ranked by spend.

Inputs: no named parameters. Returns: projects, totalCostUsd

llmkit_local_sessionA
Read-onlyIdempotent

Current session cost across all detected AI coding tools (Claude Code, Cline). No API key needed.

Inputs: no named parameters. Returns: sessions, totalCostUsd, sourceCount

llmkit_session_summaryA
Read-onlyIdempotent

Get recent proxy sessions with cost, duration, and models used

Inputs: sessionId, limit. Returns: sessions

llmkit_usage_statsC
Read-onlyIdempotent

Get usage statistics (spend, requests, top models) for a time period

Inputs: period. Returns: period, requests, totalSpendUsd, inputTokens, +4 more

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.7/5.0

    Scored across 11 tools

    Disambiguation4/5

    Tools are mostly distinct, targeting different aspects like budget, cost breakdown, health, keys, local session, etc. However, llmkit_cost_query, llmkit_usage_stats, and llmkit_session_summary have overlapping purposes (cost/spend/usage data) which could cause some confusion.

    Naming Consistency5/5

    All tools follow a consistent pattern with the 'llmkit_' prefix and descriptive noun phrases separated by underscores, e.g., llmkit_budget_status, llmkit_cost_query. No mixing of conventions.

    Tool Count5/5

    With 11 tools, the count is well within the recommended 3-15 range. Each tool covers a distinct aspect of proxy cost monitoring and management, justifying its inclusion.

    Completeness4/5

    The tool surface covers major monitoring needs: health, budget, cost, keys, sessions, and local attribution. However, it lacks management capabilities (e.g., create/update keys or budgets) and could include actions like setting usage limits, which are minor gaps for a read-only analytics server.

    Maintenance

    ActivityNo data
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      C
      quality
      D
      maintenance
      Enables interaction with Google Cloud services including billing cost analysis, log querying, and metrics monitoring through natural language commands. Provides comprehensive tools for managing GCP resources, analyzing costs, detecting anomalies, and retrieving operational insights.
      40
      1
      Apache 2.0
    • A
      license
      B
      quality
      D
      maintenance
      Enables querying Langfuse analytics, cost metrics, and usage data across multiple projects. Provides tools for trace analysis, model/service cost breakdowns, and daily usage trends through natural language queries.
      24
      73 npm
      MIT
    • F
      license
      Not graded
      quality
      D
      maintenance
      Enables querying and analyzing AWS cost and usage data through the AWS Cost Explorer API. Supports cost comparisons, forecasting, dimension filtering, and cost change driver analysis with streamable HTTP deployment to Amazon Bedrock Agentcore Runtime.
      -
    • A
      license
      B
      quality
      C
      maintenance
      An MCP server for unified cost tracking and analysis across AWS, OpenAI, and Anthropic. It enables users to query expenditures, compare costs across providers, and analyze usage trends through natural language.
      10
      MIT