Un servidor de Protocolo de Contexto Modelo (MCP) que le permite interactuar con sus transacciones y presupuestos de Lunchmoney a través de Claude y otros asistentes de IA.
Esta herramienta te permite conectar tus datos financieros de Lunchmoney con Claude AI, para que puedas hacer preguntas sobre tus gastos, analizar tu presupuesto y obtener información sobre tus finanzas a través de una conversación natural.
Related MCP server: Lunch Money MCP Server
Este servidor proporciona cuatro herramientas principales:
get-recent-transactions : Ver sus transacciones recientes de los últimos N días
buscar-transacciones : buscar transacciones por palabra clave en nombres de beneficiarios o notas
get-category-spending : Analiza el gasto en categorías específicas
get-budget-summary : Obtenga información detallada del presupuesto, incluidos gastos, montos restantes y elementos recurrentes
Importante: MCP proporciona una forma estructurada para que Claude interactúe con sus datos de Lunchmoney, respetando la privacidad. Esto es lo que debe saber:
Claude (el anfitrión) crea un cliente que se conecta a su servidor MCP local
Su token API de Lunchmoney permanece en su máquina local
El servidor MCP se ejecuta localmente y obtiene datos de la API de Lunchmoney
Se le pedirá que apruebe cada solicitud para acceder a sus datos de Lunchmoney
Cuando haces una pregunta sobre tus finanzas, Claude solicita información específica del servidor MCP
El servidor MCP procesa su solicitud localmente y devuelve solo los resultados relevantes
Claude nunca tiene acceso directo a sus datos financieros completos ni a su token API
Solo se comparte con Claude la información específica solicitada (como resúmenes de transacciones o estado del presupuesto).
Las políticas de retención de datos de Anthropic se aplican a estos resultados resumidos que forman parte de su conversación
Cada conexión de servidor está aislada, manteniendo límites de seguridad claros
Puede encontrar más información sobre MCP en la documentación: https://modelcontextprotocol.io/introduction
Consulte también la documentación oficial de Claude: https://modelcontextprotocol.io/quickstart/user
Node.js es una plataforma de software que le permite ejecutar código JavaScript en su computadora (fuera de un navegador web).
Para instalar Node.js:
Windows/Mac : Descargue y ejecute el instalador desde el sitio web oficial de Node.js
Mac con Homebrew : Ejecute brew install node en la terminal
Linux : use su administrador de paquetes (por ejemplo, sudo apt install nodejs para Ubuntu)
Una vez que Node.js esté instalado en su computadora, puede ejecutar el servidor directamente sin descargar nada:
Obtén tu token de API de Lunchmoney desde la configuración de desarrollador de Lunchmoney
Abra Claude Desktop
Vaya a Configuración → Desarrollador -> Edit Config
Agregue la siguiente configuración:
{
"mcpServers": {
"lunchmoney": {
"command": "npx",
"args": ["-y", "lunchmoney-mcp-server"],
"env": {
"LUNCHMONEY_TOKEN": "your_token_here"
}
}
}
}Reemplace your_token_here con su token API de Lunchmoney real.
Nota importante: después de cambiar la configuración, es posible que deba reiniciar Claude Desktop para que los cambios surtan efecto.
Una vez configurado en Claude Desktop, podrás hacer preguntas como:
"Muéstrame mis transacciones recientes de la semana pasada"
Buscar todas las transacciones en Amazon
"¿Cuánto gasté en restaurantes el mes pasado?"
"Buscar transacciones etiquetadas como gastos comerciales"
"Muéstrame el resumen de mi presupuesto de este mes"
¿Cuál es el estado de mi presupuesto de enero a marzo de 2024?
"¿Cuánto me queda de mi presupuesto de alimentación?"
"Muéstrame las categorías en las que me he pasado del presupuesto"
El Protocolo de Contexto de Modelo (MCP) es un protocolo abierto que estandariza cómo las aplicaciones proporcionan contexto a los Modelos de Lenguaje Grandes (LLM). Piense en MCP como un puerto USB-C para aplicaciones de IA: proporciona una forma estandarizada de conectar los modelos de IA a diferentes fuentes de datos y herramientas.
Algunos beneficios clave del MCP:
Forma estandarizada de exponer datos y funcionalidades a los LLM
Seguridad con intervención humana (todas las acciones requieren la aprobación del usuario)
Ecosistema en crecimiento de integraciones prediseñadas
Funciona con múltiples modelos y aplicaciones de IA.
Claude dice que no puede conectarse a mi servidor MCP:
Asegúrese de que la configuración en la configuración de desarrollador de Claude sea correcta
Intente reiniciar Claude Desktop después de cambiar la configuración
Comprueba que tu token API de Lunchmoney sea válido
Claude no reconoce los comandos de Lunchmoney:
Iniciar una nueva conversación en Claude
Intente mencionar explícitamente Lunchmoney en su consulta (por ejemplo, "Muéstrame mis transacciones recientes de Lunchmoney")
Los datos presupuestarios deben utilizar límites mensuales para las fechas (por ejemplo, del 01/01/2024 al 31/01/2024)
Las transacciones pueden utilizar cualquier rango de fechas
Todos los valores monetarios se devuelven en su moneda original.
Instituto Tecnológico de Massachusetts (MIT)
¡Agradecemos sus contribuciones! No dude en enviar una solicitud de incorporación de cambios.
Get budget summary for a specific time period
| Name | Required | Description | Default |
|---|---|---|---|
| start_date | Yes | Start date (YYYY-MM-DD, should be start of month) | |
| end_date | Yes | End date (YYYY-MM-DD, should be end of month) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It states the tool retrieves a 'budget summary' but doesn't explain what that summary includes (e.g., categories, totals, trends), whether it's read-only (implied but not explicit), or any limitations like rate limits or authentication needs. This leaves significant gaps for an 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 a single, efficient sentence that directly states the tool's purpose without unnecessary words. It's appropriately sized and front-loaded, making it easy to parse quickly.
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 lack of annotations and output schema, the description is incomplete. It doesn't explain what the 'budget summary' output contains, how it's structured, or any behavioral traits like error handling. For a tool with two parameters and no structured output documentation, this leaves the agent with insufficient context to use it effectively.
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 description coverage is 100%, with both parameters ('start_date' and 'end_date') fully documented in the schema. The description adds minimal value by mentioning 'time period,' which aligns with the schema but doesn't provide additional context like format examples or edge cases beyond what's already specified.
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 ('Get') and resource ('budget summary') with a specific scope ('for a specific time period'), which makes the purpose understandable. However, it doesn't differentiate this tool from its siblings like 'get-category-spending' or 'get-recent-transactions', which likely provide related but different budget/transaction 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?
The description provides no guidance on when to use this tool versus alternatives. It doesn't mention sibling tools like 'get-category-spending' or 'get-recent-transactions', nor does it specify prerequisites, exclusions, or appropriate contexts for usage beyond the time period requirement.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get spending in a category
| Name | Required | Description | Default |
|---|---|---|---|
| category | Yes | Category name | |
| days | No | Number of days to look back |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. 'Get spending' implies a read operation, but it doesn't specify if this requires authentication, has rate limits, returns historical or real-time data, or what format the spending data is in. This is a significant gap for a tool with no annotation coverage.
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, efficient sentence with zero waste. It's appropriately sized and front-loaded, making it easy to parse quickly.
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 complexity (a read operation with two parameters) and the lack of annotations and output schema, the description is incomplete. It doesn't explain what 'spending' entails (e.g., total amount, list of transactions), how results are returned, or any behavioral traits, leaving the agent with insufficient context.
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 description doesn't add any parameter semantics beyond what the schema provides. With 100% schema description coverage, the schema already documents 'category' as 'Category name' and 'days' as 'Number of days to look back' with a default of 30. The baseline score of 3 is appropriate since the schema does the heavy lifting.
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 'Get spending in a category' clearly states the verb 'Get' and resource 'spending in a category', making the purpose understandable. However, it doesn't distinguish this tool from siblings like 'get-budget-summary' or 'get-recent-transactions' which might also involve spending data, so it lacks sibling differentiation.
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 no guidance on when to use this tool versus alternatives. It doesn't mention siblings like 'search-transactions' for broader queries or 'get-budget-summary' for aggregated data, leaving the agent without context for tool selection.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get recent transactions
| Name | Required | Description | Default |
|---|---|---|---|
| days | No | Number of days to look back | |
| limit | No | Maximum number of transactions to return |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries full burden for behavioral disclosure. 'Get recent transactions' implies a read operation but reveals nothing about permissions, rate limits, pagination, error conditions, or what 'recent' means contextually. For a tool with zero annotation coverage, this leaves critical behavioral traits undocumented.
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 extremely concise at three words with zero wasted text. It's front-loaded and efficiently communicates the core function without unnecessary elaboration. While under-specified, it earns full marks for brevity and structure.
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 moderate complexity (2 parameters, no output schema, no annotations), the description is incomplete. It fails to explain what 'recent' means, how results are ordered, what data fields are returned, or how it differs from sibling tools. The agent lacks sufficient context to use this tool effectively without trial and error.
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 both parameters ('days' and 'limit') well-documented in the schema. The description adds no parameter semantics beyond what the schema already provides. According to scoring rules, when schema coverage is high (>80%), the baseline is 3 even with no param info in the description.
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 'Get recent transactions' is a tautology that essentially restates the tool name without adding meaningful differentiation. It specifies the verb 'Get' and resource 'transactions' but lacks specificity about scope or how it differs from sibling tools like 'search-transactions'. This provides minimal value beyond the name itself.
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 no guidance on when to use this tool versus alternatives like 'search-transactions' or 'get-category-spending'. There's no mention of context, prerequisites, or exclusions. The agent must infer usage from the name alone, which is insufficient for informed tool selection.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Search transactions by keyword
| Name | Required | Description | Default |
|---|---|---|---|
| keyword | Yes | Search term to look for | |
| days | No | Number of days to look back | |
| limit | No | Maximum number of transactions to return |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. While 'search' implies a read operation, the description doesn't address important behavioral aspects like whether this requires authentication, what happens with no results, whether results are paginated, or any rate limits. For a search tool with zero annotation coverage, this leaves significant behavioral questions unanswered.
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 extremely concise at just three words, with zero wasted language. It's front-loaded with the essential information (search transactions) and specifies the mechanism (by keyword) efficiently. Every word serves a clear purpose in this minimal description.
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 search tool with 3 parameters, no annotations, and no output schema, the description is inadequate. It doesn't explain what constitutes a 'transaction' in this context, what fields are searched, the format of results, or how the search algorithm works. The agent would need to guess about the tool's behavior and output based on minimal 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 description mentions searching 'by keyword' which aligns with one of the three parameters. However, with 100% schema description coverage, all parameters are already well-documented in the schema itself. The description adds minimal value beyond what's in the structured schema, meeting the baseline expectation when schema coverage is complete.
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's purpose with a specific verb ('search') and resource ('transactions'), and specifies the search mechanism ('by keyword'). However, it doesn't differentiate this tool from its sibling 'get-recent-transactions' which also deals with transactions, leaving some ambiguity about when to use one versus the other.
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 no guidance on when to use this tool versus alternatives. It doesn't mention the sibling tools (get-budget-summary, get-category-spending, get-recent-transactions) or explain when keyword searching is preferable to other transaction retrieval methods. The agent receives no contextual usage information.
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.
v1.0.0get-budget-summaryget-category-spendingget-recent-transactionssearch-transactionsScored across 4 tools
Each tool has a clearly distinct purpose: get-budget-summary targets budget overviews, get-category-spending focuses on category-level data, get-recent-transactions retrieves chronological transaction lists, and search-transactions enables keyword-based queries. There is no overlap in functionality, making tool selection straightforward for an agent.
All tool names follow a consistent verb-noun pattern with hyphens (e.g., get-budget-summary, search-transactions). The naming is uniform across all four tools, using 'get' for retrieval operations and 'search' for querying, which enhances predictability and readability.
With 4 tools, the count is reasonable for a personal finance domain, covering key operations like summaries, category analysis, and transaction retrieval. However, it feels slightly thin as it lacks tools for creating or updating data (e.g., adding transactions or categories), which might be expected in a full-featured finance server.
The tool set covers read-only operations well, including summaries, category spending, and transaction searches, but there are notable gaps. It lacks CRUD capabilities for transactions, categories, or budgets (e.g., create, update, delete), which limits agents to viewing data without modifying it, potentially causing workflow dead ends.
The Ramp MCP server enables users to securely connect Ramp with AI assistants like ChatGPT and Claude to query financial data and take actions using natural language. It transforms Ramp's developer API into a SQL interface that LLMs can query, allowing admins to analyze spend trends, identify cost savings, and run complex SQL analyses on comprehensive datasets (transactions, purchase orders, vendors, users), while all users can manage cards, view transactions, request reimbursements, and get expense policy answers.
An MCP server that provides read access to your cloud storage providers, bank accounts and more.
Query your real net worth, spending, transactions, budgets and portfolio from any MCP client.
Personal finance for AI agents — onboard, import statements, categorize & budget over MCP.
