[{"_1":2,"_296":-5,"_297":-5},"loaderData",{"_3":4,"_29":30,"_31":32,"_244":245},"root",{"_5":6,"_14":15},"toasts",[7],{"_8":9,"_10":11,"_12":13},"uid","019726c0-ed2b-752f-bf05-f6c328d70f20","title","This MCP server does not support inspection","type","error","visitor",{"_16":17,"_18":19},"viewportWidth",1280,"visitorSession",{"_20":21,"_23":24,"_25":-5,"_26":27,"_28":-5},"attributes",[22],"BOT","id",7058884,"membership","referenceId","01971406-43e3-74e8-b880-2f000258007a","userAccount","routes/_public/_layout",{},"routes/_public/mcp/servers/~namespace/~slug/_pages/_layout",{"_33":34,"_75":-5,"_76":35,"_36":77,"_162":163,"_243":138},"availableLocales",[35,47,54,61,68],{"_36":37,"_40":44,"_45":46},"mcpServer",{"_38":39,"_40":41,"_42":43},"description","Provides integration between Merge API and LLM providers supporting the MCP protocol, allowing natural language interaction with Merge data across HRIS, ATS, and other categories.","name","Merge MCP Server","readme","# Merge MCP Server\n\nThis MCP (Model Context Protocol) server provides integration between Merge API and any LLM provider supporting the MCP protocol (e.g., Claude for Desktop), allowing you to interact with your Merge data using natural language.\n\n## ✨ Features\n- Query Merge API entities using natural language\n- Get information about your Merge data models and their fields\n- Create and update entities through conversational interfaces\n- Support for multiple Merge API categories (HRIS, ATS, etc.)\n\n## 📦 Installation\n\n### Prerequisites\n\n- A Merge API key and account token\n- Python 3.10 or higher\n- [uv](https://github.com/astral-sh/uv)\n\nInstall `uv` with standalone installer:\n```bash\n# On macOS and Linux.\ncurl -LsSf https://astral.sh/uv/install.sh | sh\n\n# On Windows.\npowershell -ExecutionPolicy ByPass -c \"irm https://astral.sh/uv/install.ps1 | iex\"\n```\n\nor through pip:\n```bash\n# With pip.\npip install uv\n\n# With pipx.\npipx install uv\n```\n\n## 🔌 MCP setup\nHere is an example config file which you can use to set up Merge MCP.\n\n```json\n{\n \"mcpServers\": {\n \"merge-mcp-server\": {\n \"command\": \"uvx\",\n \"args\": [\"merge-mcp\"],\n \"env\": {\n \"MERGE_API_KEY\": \"your_api_key\",\n \"MERGE_ACCOUNT_TOKEN\": \"your_account_token\"\n }\n }\n }\n}\n```\nNote: If \"uvx\" command does not work, try absolute path (i.e. /Users/username/.local/bin/uvx)\n\n### Example Claude Desktop configuration\n\n1. Ensure you have `uvx` installed\n\n2. Download [Claude Desktop](https://claude.ai/download) from the official website\n\n3. Once downloaded, open the app and follow the instructions to set up your account\n\n4. Navigate to **Settings → Developer → Edit Config**. This should open a file named `claude_desktop_config.json` in a text editor.\n\n5. Copy the MCP server setup JSON above and paste it into the text editor\n\n6. Replace `your_api_key` and `your_account_token` with your actual Merge API key and Linked Account token. You will also need to replace `uvx` with the absolute path to the command in the config file (i.e. `/Users/username/.local/bin/uvx`). You can find the absolute path by running `which uvx` through your terminal.\n\n7. Save the config file\n\n8. Restart Claude Desktop to see your tools. The tools may take a minute to appear\n\n### Example Python client configuration\n\n1. Setting up your environment\n\n```bash\n# Create project directory\nmkdir mcp-client\ncd mcp-client\n\n# Create virtual environment\npython -m venv .venv\n\n# Activate virtual environment\n# On Windows:\n.venv\\Scripts\\activate\n# On Unix or MacOS:\nsource .venv/bin/activate\n\n# Install required packages\npip install mcp uv anthropic python-dotenv\n\n# Create our main file\ntouch client.py\n```\n\n2. Setting up your API keys\n\n```bash\n# Add your ANTHROPIC_API_KEY and MERGE_API_KEY to .env\necho \"ANTHROPIC_API_KEY=\" >> .env\necho \"MERGE_API_KEY=\" >> .env\n\n# Add .env file to .gitignore\necho \".env\" >> .gitignore\n```\n\n3. Create a `client.py` file and add the following code\n\n```python\nimport os\nimport asyncio\nfrom typing import Optional\nfrom contextlib import AsyncExitStack\n\nfrom mcp import ClientSession, StdioServerParameters\nfrom mcp.client.stdio import stdio_client\n\nfrom anthropic import Anthropic\nfrom dotenv import load_dotenv\n\nload_dotenv() # load environment variables from .env\n\nclass MCPClient:\n def __init__(self):\n # Initialize session and client objects\n self.session: Optional[ClientSession] = None\n self.exit_stack = AsyncExitStack()\n self.anthropic = Anthropic()\n\n # Methods will go here\n```\n\n4. Add a `connect_to_server` function to the MCPClient class\n\n```python\nasync def connect_to_server(self, linked_account_token: str):\n \"\"\"Connect to an MCP server\n Args:\n linked_account_token: The token for the associated Linked Account\n \"\"\"\n\n server_params = StdioServerParameters(\n command=\"uvx\",\n args=[\"merge-mcp\"],\n env={\n \"MERGE_API_KEY\": os.getenv(\"MERGE_API_KEY\"),\n \"MERGE_ACCOUNT_TOKEN\": linked_account_token\n }\n )\n\n stdio_transport = await self.exit_stack.enter_async_context(stdio_client(server_params))\n self.stdio, self.write = stdio_transport\n self.session = await self.exit_stack.enter_async_context(ClientSession(self.stdio, self.write))\n\n await self.session.initialize()\n\n # List available tools\n response = await self.session.list_tools()\n tools = response.tools\n print(\"\\nConnected to server with tools:\", [tool.name for tool in tools])\n```\n\n5. Add a `process_query` function to the MCPClient class\n\n```python\nasync def process_query(self, query: str) -> str:\n \"\"\"Process a query using Claude and available tools\"\"\"\n messages = [\n {\n \"role\": \"user\",\n \"content\": query\n }\n ]\n\n response = await self.session.list_tools()\n available_tools = [{\n \"name\": tool.name,\n \"description\": tool.description,\n \"input_schema\": tool.inputSchema\n } for tool in response.tools]\n\n # Initial Claude API call\n response = self.anthropic.messages.create(\n model=\"claude-3-5-sonnet-20241022\",\n max_tokens=1000,\n messages=messages,\n tools=available_tools\n )\n\n # Process response and handle tool calls\n final_text = []\n assistant_message_content = []\n for content in response.content:\n if content.type == 'text':\n final_text.append(content.text)\n assistant_message_content.append(content)\n\n elif content.type == 'tool_use':\n tool_name = content.name\n tool_args = content.input\n\n # Get confirmation for tool call execution\n confirmation = input(f\"Do you want to call tool '{tool_name}' with arguments {tool_args}? (y/n): \").strip().lower()\n if confirmation.startswith('y'):\n result = await self.session.call_tool(tool_name, tool_args)\n final_text.append(f\"[Calling tool {tool_name} with args {tool_args}]\")\n assistant_message_content.append(content)\n messages.append({\n \"role\": \"assistant\",\n \"content\": assistant_message_content\n })\n messages.append({\n \"role\": \"user\",\n \"content\": [\n {\n \"type\": \"tool_result\",\n \"tool_use_id\": content.id,\n \"content\": result.content\n }\n ]\n })\n\n # Get next response from Claude\n response = self.anthropic.messages.create(\n model=\"claude-3-5-sonnet-20241022\",\n max_tokens=1000,\n messages=messages,\n tools=available_tools\n )\n final_text.append(response.content[0].text)\n\n else:\n final_text.append(f\"[Skipped calling tool {tool_name} with args {tool_args}]\")\n\n return \"\\n\".join(final_text)\n```\n\n6. Add a `chat_loop` function to the MCPClient class\n\n```python\n async def chat_loop(self):\n \"\"\"Run an interactive chat loop\"\"\"\n print(\"\\nMCP Client Started!\")\n print(\"Type your queries or 'quit' to exit.\")\n\n while True:\n try:\n query = input(\"\\nQuery: \").strip()\n\n if query.lower() == 'quit':\n break\n\n response = await self.process_query(query)\n print(\"\\n\" + response)\n\n except Exception as e:\n print(f\"\\nError: {str(e)}\")\n```\n\n7. Add a `cleanup` function to the MCPClient class\n\n```python\n async def cleanup(self):\n \"\"\"Clean up resources\"\"\"\n await self.exit_stack.aclose()\n```\n\n8. Add a `main` function to the `client.py` file as the main entry point\n\n```python\nasync def main():\n client = MCPClient()\n try:\n await client.connect_to_server(\"\")\n await client.chat_loop()\n finally:\n await client.cleanup()\n\nif __name__ == \"__main__\":\n import sys\n asyncio.run(main())\n```\n\n9. Running the client\n\n```bash\npython client.py\n```\n\n## 🔍 Scopes\n\nScopes determine which tools are enabled on the MCP server and are used to control access to different parts of the Merge API. **If no scopes are specified, all available scopes will be enabled.**\n\nWhen starting the server, you can specify which scopes to enable. This is done by passing the `--scopes` flag with a list of scopes.\n\n```json\n{\n \"mcpServers\": {\n \"merge-mcp-server\": {\n \"command\": \"uvx\",\n \"args\": [\n \"merge-mcp\",\n \"--scopes\",\n \"ats.Job:read\",\n \"ats.Candidate\",\n \"ats.Application:write\"\n ],\n \"env\": {\n \"MERGE_API_KEY\": \"your_api_key\",\n \"MERGE_ACCOUNT_TOKEN\": \"your_account_token\"\n }\n }\n }\n}\n```\n\n### Scope Format\n\nScopes in the Merge MCP server follow a specific format based on the Merge API category and common model names. Each scope is formatted as:\n\n```\n.:\n```\n\nWhere:\n- `` is the Merge API category (e.g., `hris`, `ats`, `accounting`)\n- `` is the name of the Merge Common Model (e.g., `Employee`, `Candidate`, `Account`)\n- `` is either `read` or `write` (optional - if not specified, all permissions are granted)\n\nExamples of valid scopes:\n- `hris.Employee:read` - Allows reading employee data from HRIS category\n- `ats.Candidate:write` - Allows creating or updating candidate data in ATS category\n- `accounting.Account` - Allows all operations on account data in Accounting category\n\nYou can combine multiple scopes to grant different permissions.\n\n### Important Notes on Scope Availability\n\nThe available scopes depend on your Merge API account configuration and the models the Linked Account has access to. Scopes must be cross-referenced with enabled scopes on your Linked Account:\n\n- **Category Mismatch**: If you specify a scope for a category that doesn't match your Linked Account (e.g., using `ats.Job` with an HRIS Linked Account), no tools for that scope will be returned.\n\n- **Permission Mismatch**: If you request a permission that isn't enabled for your Linked Account (e.g., using `hris.Employee:write` when only read access is enabled), tools requiring that permission won't be returned.\n\n- **Validation**: The server will automatically validate your requested scopes against what's available in your Linked Account and will only enable tools for valid, authorized scopes.\n\nScopes typically correspond to different models or entity types in the Merge API, and they control both read and write access to these entities.\n\n## 🚀 Available Tools\n\nThe Merge MCP server provides access to various Merge API endpoints as tools. The available tools depend on your Merge API category (HRIS, ATS, etc.) and the scopes you have enabled.\n\nTools are dynamically generated based on your Merge API schema and include operations for:\n\n- Retrieving entity details\n- Listing entities\n- Creating new entities\n- Updating existing entities\n- And more, based on your specific Merge API configuration\n\n**Note:** Download tools are not currently supported. This is a known limitation and will be addressed in a future release.\n\n## 🔑 Environment Variables\n\nThe following environment variables are used by the Merge MCP server:\n\n- `MERGE_API_KEY`: Your Merge API key\n- `MERGE_ACCOUNT_TOKEN`: Your Merge Linked Account token\n- `MERGE_TENANT` (Optional): The Merge API tenant. Valid values are `US`, `EU`, and `APAC`. Defaults to `US`.","English","tag","en-US",{"_36":48,"_40":52,"_45":53},{"_38":49,"_40":50,"_42":51},"Proporciona integración entre Merge API y proveedores LLM compatibles con el protocolo MCP, lo que permite la interacción en lenguaje natural con los datos de Merge en HRIS, ATS y otras categorías.","Fusionar servidor MCP","# Fusionar servidor MCP\n\nEste servidor MCP (Protocolo de contexto de modelo) proporciona integración entre Merge API y cualquier proveedor LLM que admita el protocolo MCP (por ejemplo, Claude for Desktop), lo que le permite interactuar con sus datos de Merge utilizando lenguaje natural.\n\n## ✨ Características\n\n* Consultar entidades de la API de combinación mediante lenguaje natural\n* Obtenga información sobre sus modelos de datos Merge y sus campos\n* Crear y actualizar entidades a través de interfaces conversacionales\n* Compatibilidad con múltiples categorías de API de combinación (HRIS, ATS, etc.)\n\n## 📦 Instalación\n\n### Prerrequisitos\n\n* Una clave API de combinación y un token de cuenta\n* Python 3.10 o superior\n* [ultravioleta](https://github.com/astral-sh/uv)\n\nInstalar `uv` con el instalador independiente:\n\n```bash\n# On macOS and Linux.\ncurl -LsSf https://astral.sh/uv/install.sh | sh\n\n# On Windows.\npowershell -ExecutionPolicy ByPass -c \"irm https://astral.sh/uv/install.ps1 | iex\"\n```\n\no a través de pip:\n\n```bash\n# With pip.\npip install uv\n\n# With pipx.\npipx install uv\n```\n\n## Configuración de MCP\n\nAquí hay un archivo de configuración de ejemplo que puedes usar para configurar Merge MCP.\n\n```json\n{\n \"mcpServers\": {\n \"merge-mcp-server\": {\n \"command\": \"uvx\",\n \"args\": [\"merge-mcp\"],\n \"env\": {\n \"MERGE_API_KEY\": \"your_api_key\",\n \"MERGE_ACCOUNT_TOKEN\": \"your_account_token\"\n }\n }\n }\n}\n```\n\nNota: Si el comando \"uvx\" no funciona, intente con la ruta absoluta (es decir, /Users/username/.local/bin/uvx)\n\n### Ejemplo de configuración de Claude Desktop\n\n1. Asegúrese de tener `uvx` instalado\n\n2. Descarga [Claude Desktop](https://claude.ai/download) desde el sitio web oficial\n\n3. Una vez descargada, abra la aplicación y siga las instrucciones para configurar su cuenta.\n\n4. Vaya a **Configuración → Desarrollador → Editar configuración** . Esto debería abrir un archivo llamado `claude_desktop_config.json` en un editor de texto.\n\n5. Copie el JSON de configuración del servidor MCP anterior y péguelo en el editor de texto\n\n6. Reemplaza `your_api_key` y `your_account_token` con tu clave de API de Merge y el token de la cuenta vinculada. También deberás reemplazar `uvx` con la ruta absoluta del comando en el archivo de configuración (es decir, `/Users/username/.local/bin/uvx` ). Puedes encontrar la ruta absoluta ejecutando `which uvx` en tu terminal.\n\n7. Guardar el archivo de configuración\n\n8. Reinicie Claude Desktop para ver sus herramientas. Las herramientas pueden tardar un minuto en aparecer.\n\n### Ejemplo de configuración de cliente Python\n\n1. Configuración de su entorno\n\n```bash\n# Create project directory\nmkdir mcp-client\ncd mcp-client\n\n# Create virtual environment\npython -m venv .venv\n\n# Activate virtual environment\n# On Windows:\n.venv\\Scripts\\activate\n# On Unix or MacOS:\nsource .venv/bin/activate\n\n# Install required packages\npip install mcp uv anthropic python-dotenv\n\n# Create our main file\ntouch client.py\n```\n\n2. Configuración de sus claves API\n\n```bash\n# Add your ANTHROPIC_API_KEY and MERGE_API_KEY to .env\necho \"ANTHROPIC_API_KEY=\" >> .env\necho \"MERGE_API_KEY=\" >> .env\n\n# Add .env file to .gitignore\necho \".env\" >> .gitignore\n```\n\n3. Cree un archivo `client.py` y agregue el siguiente código\n\n```python\nimport os\nimport asyncio\nfrom typing import Optional\nfrom contextlib import AsyncExitStack\n\nfrom mcp import ClientSession, StdioServerParameters\nfrom mcp.client.stdio import stdio_client\n\nfrom anthropic import Anthropic\nfrom dotenv import load_dotenv\n\nload_dotenv() # load environment variables from .env\n\nclass MCPClient:\n def __init__(self):\n # Initialize session and client objects\n self.session: Optional[ClientSession] = None\n self.exit_stack = AsyncExitStack()\n self.anthropic = Anthropic()\n\n # Methods will go here\n```\n\n4. Agregue una función `connect_to_server` a la clase MCPClient\n\n```python\nasync def connect_to_server(self, linked_account_token: str):\n \"\"\"Connect to an MCP server\n Args:\n linked_account_token: The token for the associated Linked Account\n \"\"\"\n\n server_params = StdioServerParameters(\n command=\"uvx\",\n args=[\"merge-mcp\"],\n env={\n \"MERGE_API_KEY\": os.getenv(\"MERGE_API_KEY\"),\n \"MERGE_ACCOUNT_TOKEN\": linked_account_token\n }\n )\n\n stdio_transport = await self.exit_stack.enter_async_context(stdio_client(server_params))\n self.stdio, self.write = stdio_transport\n self.session = await self.exit_stack.enter_async_context(ClientSession(self.stdio, self.write))\n\n await self.session.initialize()\n\n # List available tools\n response = await self.session.list_tools()\n tools = response.tools\n print(\"\\nConnected to server with tools:\", [tool.name for tool in tools])\n```\n\n5. Agregue una función `process_query` a la clase MCPClient\n\n```python\nasync def process_query(self, query: str) -> str:\n \"\"\"Process a query using Claude and available tools\"\"\"\n messages = [\n {\n \"role\": \"user\",\n \"content\": query\n }\n ]\n\n response = await self.session.list_tools()\n available_tools = [{\n \"name\": tool.name,\n \"description\": tool.description,\n \"input_schema\": tool.inputSchema\n } for tool in response.tools]\n\n # Initial Claude API call\n response = self.anthropic.messages.create(\n model=\"claude-3-5-sonnet-20241022\",\n max_tokens=1000,\n messages=messages,\n tools=available_tools\n )\n\n # Process response and handle tool calls\n final_text = []\n assistant_message_content = []\n for content in response.content:\n if content.type == 'text':\n final_text.append(content.text)\n assistant_message_content.append(content)\n\n elif content.type == 'tool_use':\n tool_name = content.name\n tool_args = content.input\n\n # Get confirmation for tool call execution\n confirmation = input(f\"Do you want to call tool '{tool_name}' with arguments {tool_args}? (y/n): \").strip().lower()\n if confirmation.startswith('y'):\n result = await self.session.call_tool(tool_name, tool_args)\n final_text.append(f\"[Calling tool {tool_name} with args {tool_args}]\")\n assistant_message_content.append(content)\n messages.append({\n \"role\": \"assistant\",\n \"content\": assistant_message_content\n })\n messages.append({\n \"role\": \"user\",\n \"content\": [\n {\n \"type\": \"tool_result\",\n \"tool_use_id\": content.id,\n \"content\": result.content\n }\n ]\n })\n\n # Get next response from Claude\n response = self.anthropic.messages.create(\n model=\"claude-3-5-sonnet-20241022\",\n max_tokens=1000,\n messages=messages,\n tools=available_tools\n )\n final_text.append(response.content[0].text)\n\n else:\n final_text.append(f\"[Skipped calling tool {tool_name} with args {tool_args}]\")\n\n return \"\\n\".join(final_text)\n```\n\n6. Agregue una función `chat_loop` a la clase MCPClient\n\n```python\n async def chat_loop(self):\n \"\"\"Run an interactive chat loop\"\"\"\n print(\"\\nMCP Client Started!\")\n print(\"Type your queries or 'quit' to exit.\")\n\n while True:\n try:\n query = input(\"\\nQuery: \").strip()\n\n if query.lower() == 'quit':\n break\n\n response = await self.process_query(query)\n print(\"\\n\" + response)\n\n except Exception as e:\n print(f\"\\nError: {str(e)}\")\n```\n\n7. Agregue una función `cleanup` a la clase MCPClient\n\n```python\n async def cleanup(self):\n \"\"\"Clean up resources\"\"\"\n await self.exit_stack.aclose()\n```\n\n8. Agregue una función `main` al archivo `client.py` como punto de entrada principal\n\n```python\nasync def main():\n client = MCPClient()\n try:\n await client.connect_to_server(\"\")\n await client.chat_loop()\n finally:\n await client.cleanup()\n\nif __name__ == \"__main__\":\n import sys\n asyncio.run(main())\n```\n\n9. Ejecutando el cliente\n\n```bash\npython client.py\n```\n\n## 🔍 Ámbitos\n\nLos ámbitos determinan qué herramientas están habilitadas en el servidor MCP y se utilizan para controlar el acceso a diferentes partes de la API Merge. **Si no se especifica ningún ámbito, se habilitarán todos los ámbitos disponibles.**\n\nAl iniciar el servidor, puede especificar los ámbitos que desea habilitar. Esto se hace pasando el indicador `--scopes` con una lista de ámbitos.\n\n```json\n{\n \"mcpServers\": {\n \"merge-mcp-server\": {\n \"command\": \"uvx\",\n \"args\": [\n \"merge-mcp\",\n \"--scopes\",\n \"ats.Job:read\",\n \"ats.Candidate\",\n \"ats.Application:write\"\n ],\n \"env\": {\n \"MERGE_API_KEY\": \"your_api_key\",\n \"MERGE_ACCOUNT_TOKEN\": \"your_account_token\"\n }\n }\n }\n}\n```\n\n### Formato de alcance\n\nLos ámbitos en el servidor Merge MCP siguen un formato específico basado en la categoría de la API Merge y los nombres de modelos comunes. Cada ámbito tiene el siguiente formato:\n\n```\n.:\n```\n\nDónde:\n\n* `` es la categoría de la API de combinación (por ejemplo, `hris` , `ats` , `accounting` )\n* `` es el nombre del modelo común de fusión (por ejemplo, `Employee` , `Candidate` , `Account` )\n* `` es `read` o `write` (opcional; si no se especifica, se conceden todos los permisos)\n\nEjemplos de ámbitos válidos:\n\n* `hris.Employee:read` : permite leer datos de empleados de la categoría HRIS\n* `ats.Candidate:write` - Permite crear o actualizar datos de candidatos en la categoría ATS\n* `accounting.Account` : permite todas las operaciones sobre datos de cuentas en la categoría Contabilidad\n\nPuede combinar varios ámbitos para otorgar diferentes permisos.\n\n### Notas importantes sobre la disponibilidad del alcance\n\nLos ámbitos disponibles dependen de la configuración de su cuenta de Merge API y de los modelos a los que la cuenta vinculada tiene acceso. Los ámbitos deben tener referencias cruzadas con los ámbitos habilitados en su cuenta vinculada:\n\n* **No coincide con la categoría** : si especifica un alcance para una categoría que no coincide con su cuenta vinculada (por ejemplo, si utiliza `ats.Job` con una cuenta vinculada HRIS), no se devolverán herramientas para ese alcance.\n\n* **Falta de coincidencia de permisos** : si solicita un permiso que no está habilitado para su cuenta vinculada (por ejemplo, usar `hris.Employee:write` cuando solo está habilitado el acceso de lectura), no se devolverán las herramientas que requieran ese permiso.\n\n* **Validación** : el servidor validará automáticamente los alcances solicitados con lo que esté disponible en su cuenta vinculada y solo habilitará herramientas para alcances válidos y autorizados.\n\nLos alcances generalmente corresponden a diferentes modelos o tipos de entidades en la API de combinación y controlan el acceso de lectura y escritura a estas entidades.\n\n## 🚀 Herramientas disponibles\n\nEl servidor MCP de Merge proporciona acceso a varios puntos finales de la API de Merge como herramientas. Las herramientas disponibles dependen de la categoría de la API de Merge (HRIS, ATS, etc.) y de los ámbitos habilitados.\n\nLas herramientas se generan dinámicamente en función de su esquema de API de combinación e incluyen operaciones para:\n\n* Recuperando detalles de la entidad\n* Entidades de cotización\n* Creando nuevas entidades\n* Actualización de entidades existentes\n* Y más, según su configuración específica de Merge API\n\n**Nota:** Las herramientas de descarga no son compatibles actualmente. Esta es una limitación conocida y se solucionará en una próxima versión.\n\n## 🔑 Variables de entorno\n\nEl servidor Merge MCP utiliza las siguientes variables de entorno:\n\n* `MERGE_API_KEY` : Su clave API de Merge\n* `MERGE_ACCOUNT_TOKEN` : Su token de cuenta vinculada a la fusión\n* `MERGE_TENANT` (Opcional): El inquilino de la API Merge. Los valores válidos son `US` , `EU` y `APAC` . El valor predeterminado es `US` .","Spanish","es-ES",{"_36":55,"_40":59,"_45":60},{"_38":56,"_40":57,"_42":58},"Merge API と MCP プロトコルをサポートする LLM プロバイダー間の統合を提供し、HRIS、ATS、その他のカテゴリにわたる Merge データとの自然言語による対話を可能にします。","MCPサーバーの統合","# MCPサーバーの統合\n\nこの MCP (モデルコンテキストプロトコル) サーバーは、Merge API と MCP プロトコルをサポートする任意の LLM プロバイダー (例: Claude for Desktop) との統合を提供し、自然言語を使用して Merge データを操作できるようにします。\n\n## ✨ 特徴\n\n* 自然言語を使用して Merge API エンティティをクエリする\n* Mergeデータモデルとそのフィールドに関する情報を取得します\n* 会話型インターフェースを通じてエンティティを作成および更新する\n* 複数の Merge API カテゴリ (HRIS、ATS など) のサポート\n\n## 📦 インストール\n\n### 前提条件\n\n* Merge APIキーとアカウントトークン\n* Python 3.10以上\n* [紫外線](https://github.com/astral-sh/uv)\n\nスタンドアロンインストーラーで`uv`をインストールします。\n\n```bash\n# On macOS and Linux.\ncurl -LsSf https://astral.sh/uv/install.sh | sh\n\n# On Windows.\npowershell -ExecutionPolicy ByPass -c \"irm https://astral.sh/uv/install.ps1 | iex\"\n```\n\nまたはpip経由:\n\n```bash\n# With pip.\npip install uv\n\n# With pipx.\npipx install uv\n```\n\n## 🔌 MCP セットアップ\n\nMerge MCP を設定するために使用できる設定ファイルの例を次に示します。\n\n```json\n{\n \"mcpServers\": {\n \"merge-mcp-server\": {\n \"command\": \"uvx\",\n \"args\": [\"merge-mcp\"],\n \"env\": {\n \"MERGE_API_KEY\": \"your_api_key\",\n \"MERGE_ACCOUNT_TOKEN\": \"your_account_token\"\n }\n }\n }\n}\n```\n\n注: 「uvx」コマンドが機能しない場合は、絶対パス（例: /Users/username/.local/bin/uvx）を試してください。\n\n### Claude Desktop の構成例\n\n1. `uvx`がインストールされていることを確認してください\n\n2. 公式サイトから[Claude Desktop](https://claude.ai/download)をダウンロードしてください\n\n3. ダウンロードしたら、アプリを開いて指示に従ってアカウントを設定してください。\n\n4. \\*\\*「設定」→「開発者」→「設定の編集」\\*\\*に移動します。テキストエディターで`claude_desktop_config.json`というファイルが開きます。\n\n5. 上記のMCPサーバーセットアップJSONをコピーしてテキストエディターに貼り付けます。\n\n6. `your_api_key`と`your_account_token`を、実際の Merge API キーとリンク済みアカウントトークンに置き換えてください。また、 `uvx`設定ファイル内のコマンドへの絶対パス（例: `/Users/username/.local/bin/uvx` ）に置き換える必要があります。絶対パスは、ターミナルで`which uvx`実行することで確認できます。\n\n7. 設定ファイルを保存する\n\n8. ツールを表示するには、Claude Desktopを再起動してください。ツールが表示されるまで1分ほどかかる場合があります。\n\n### Pythonクライアントの設定例\n\n1. 環境の設定\n\n```bash\n# Create project directory\nmkdir mcp-client\ncd mcp-client\n\n# Create virtual environment\npython -m venv .venv\n\n# Activate virtual environment\n# On Windows:\n.venv\\Scripts\\activate\n# On Unix or MacOS:\nsource .venv/bin/activate\n\n# Install required packages\npip install mcp uv anthropic python-dotenv\n\n# Create our main file\ntouch client.py\n```\n\n2. APIキーの設定\n\n```bash\n# Add your ANTHROPIC_API_KEY and MERGE_API_KEY to .env\necho \"ANTHROPIC_API_KEY=\" >> .env\necho \"MERGE_API_KEY=\" >> .env\n\n# Add .env file to .gitignore\necho \".env\" >> .gitignore\n```\n\n3. `client.py`ファイルを作成し、次のコードを追加します。\n\n```python\nimport os\nimport asyncio\nfrom typing import Optional\nfrom contextlib import AsyncExitStack\n\nfrom mcp import ClientSession, StdioServerParameters\nfrom mcp.client.stdio import stdio_client\n\nfrom anthropic import Anthropic\nfrom dotenv import load_dotenv\n\nload_dotenv() # load environment variables from .env\n\nclass MCPClient:\n def __init__(self):\n # Initialize session and client objects\n self.session: Optional[ClientSession] = None\n self.exit_stack = AsyncExitStack()\n self.anthropic = Anthropic()\n\n # Methods will go here\n```\n\n4. MCPClientクラスに`connect_to_server`関数を追加する\n\n```python\nasync def connect_to_server(self, linked_account_token: str):\n \"\"\"Connect to an MCP server\n Args:\n linked_account_token: The token for the associated Linked Account\n \"\"\"\n\n server_params = StdioServerParameters(\n command=\"uvx\",\n args=[\"merge-mcp\"],\n env={\n \"MERGE_API_KEY\": os.getenv(\"MERGE_API_KEY\"),\n \"MERGE_ACCOUNT_TOKEN\": linked_account_token\n }\n )\n\n stdio_transport = await self.exit_stack.enter_async_context(stdio_client(server_params))\n self.stdio, self.write = stdio_transport\n self.session = await self.exit_stack.enter_async_context(ClientSession(self.stdio, self.write))\n\n await self.session.initialize()\n\n # List available tools\n response = await self.session.list_tools()\n tools = response.tools\n print(\"\\nConnected to server with tools:\", [tool.name for tool in tools])\n```\n\n5. MCPClientクラスに`process_query`関数を追加する\n\n```python\nasync def process_query(self, query: str) -> str:\n \"\"\"Process a query using Claude and available tools\"\"\"\n messages = [\n {\n \"role\": \"user\",\n \"content\": query\n }\n ]\n\n response = await self.session.list_tools()\n available_tools = [{\n \"name\": tool.name,\n \"description\": tool.description,\n \"input_schema\": tool.inputSchema\n } for tool in response.tools]\n\n # Initial Claude API call\n response = self.anthropic.messages.create(\n model=\"claude-3-5-sonnet-20241022\",\n max_tokens=1000,\n messages=messages,\n tools=available_tools\n )\n\n # Process response and handle tool calls\n final_text = []\n assistant_message_content = []\n for content in response.content:\n if content.type == 'text':\n final_text.append(content.text)\n assistant_message_content.append(content)\n\n elif content.type == 'tool_use':\n tool_name = content.name\n tool_args = content.input\n\n # Get confirmation for tool call execution\n confirmation = input(f\"Do you want to call tool '{tool_name}' with arguments {tool_args}? (y/n): \").strip().lower()\n if confirmation.startswith('y'):\n result = await self.session.call_tool(tool_name, tool_args)\n final_text.append(f\"[Calling tool {tool_name} with args {tool_args}]\")\n assistant_message_content.append(content)\n messages.append({\n \"role\": \"assistant\",\n \"content\": assistant_message_content\n })\n messages.append({\n \"role\": \"user\",\n \"content\": [\n {\n \"type\": \"tool_result\",\n \"tool_use_id\": content.id,\n \"content\": result.content\n }\n ]\n })\n\n # Get next response from Claude\n response = self.anthropic.messages.create(\n model=\"claude-3-5-sonnet-20241022\",\n max_tokens=1000,\n messages=messages,\n tools=available_tools\n )\n final_text.append(response.content[0].text)\n\n else:\n final_text.append(f\"[Skipped calling tool {tool_name} with args {tool_args}]\")\n\n return \"\\n\".join(final_text)\n```\n\n6. MCPClientクラスに`chat_loop`関数を追加する\n\n```python\n async def chat_loop(self):\n \"\"\"Run an interactive chat loop\"\"\"\n print(\"\\nMCP Client Started!\")\n print(\"Type your queries or 'quit' to exit.\")\n\n while True:\n try:\n query = input(\"\\nQuery: \").strip()\n\n if query.lower() == 'quit':\n break\n\n response = await self.process_query(query)\n print(\"\\n\" + response)\n\n except Exception as e:\n print(f\"\\nError: {str(e)}\")\n```\n\n7. MCPClientクラスに`cleanup`関数を追加する\n\n```python\n async def cleanup(self):\n \"\"\"Clean up resources\"\"\"\n await self.exit_stack.aclose()\n```\n\n8. `client.py`ファイルにメインエントリポイントとして`main`関数を追加します。\n\n```python\nasync def main():\n client = MCPClient()\n try:\n await client.connect_to_server(\"\")\n await client.chat_loop()\n finally:\n await client.cleanup()\n\nif __name__ == \"__main__\":\n import sys\n asyncio.run(main())\n```\n\n9. クライアントの実行\n\n```bash\npython client.py\n```\n\n## 🔍 スコープ\n\nスコープは、MCPサーバー上で有効にするツールを決定し、Merge APIの様々な部分へのアクセスを制御するために使用されます。**スコープが指定されていない場合は、利用可能なすべてのスコープが有効になります。**\n\nサーバーを起動する際に、有効にするスコープを指定できます。これは、スコープのリストを指定した`--scopes`フラグを渡すことで実行されます。\n\n```json\n{\n \"mcpServers\": {\n \"merge-mcp-server\": {\n \"command\": \"uvx\",\n \"args\": [\n \"merge-mcp\",\n \"--scopes\",\n \"ats.Job:read\",\n \"ats.Candidate\",\n \"ats.Application:write\"\n ],\n \"env\": {\n \"MERGE_API_KEY\": \"your_api_key\",\n \"MERGE_ACCOUNT_TOKEN\": \"your_account_token\"\n }\n }\n }\n}\n```\n\n### スコープ形式\n\nMerge MCPサーバーのスコープは、Merge APIのカテゴリと一般的なモデル名に基づいた特定の形式に従います。各スコープの形式は以下のとおりです。\n\n```\n.:\n```\n\nどこ：\n\n* ``は Merge API カテゴリです (例: `hris` 、 `ats` 、 `accounting` )\n* ``は、マージ共通モデルの名前です (例: `Employee` 、 `Candidate` 、 `Account` )\n* ``は`read`または`write`いずれかです (オプション - 指定されていない場合はすべての権限が付与されます)\n\n有効なスコープの例:\n\n* `hris.Employee:read` - HRIS カテゴリから従業員データを読み取ることができます\n* `ats.Candidate:write` - ATS カテゴリの候補者データの作成または更新を許可します\n* `accounting.Account` - 会計カテゴリのアカウントデータに対するすべての操作を許可します\n\n複数のスコープを組み合わせて、異なる権限を付与することができます。\n\n### スコープの可用性に関する重要な注意事項\n\n利用可能なスコープは、Merge APIアカウントの設定と、リンクされたアカウントがアクセスできるモデルによって異なります。スコープは、リンクされたアカウントで有効なスコープと相互参照する必要があります。\n\n* **カテゴリの不一致**: リンクされたアカウントと一致しないカテゴリのスコープを指定した場合 (例: HRIS リンクされたアカウントで`ats.Job`を使用する)、そのスコープのツールは返されません。\n\n* **権限の不一致**: リンクされたアカウントで有効になっていない権限をリクエストした場合 (読み取りアクセスのみが有効になっている場合に`hris.Employee:write`を使用するなど)、その権限を必要とするツールは返されません。\n\n* **検証**: サーバーは、要求されたスコープをリンクされたアカウントで利用可能なスコープと照らし合わせて自動的に検証し、有効で承認されたスコープに対してのみツールを有効にします。\n\nスコープは通常、Merge API 内のさまざまなモデルまたはエンティティタイプに対応し、これらのエンティティへの読み取りアクセスと書き込みアクセスの両方を制御します。\n\n## 🚀 利用可能なツール\n\nMerge MCPサーバーは、様々なMerge APIエンドポイントへのアクセスをツールとして提供します。利用可能なツールは、Merge APIのカテゴリー（HRIS、ATSなど）と有効にしたスコープによって異なります。\n\nツールは Merge API スキーマに基づいて動的に生成され、次の操作が含まれます。\n\n* エンティティの詳細を取得しています\n* エンティティのリスト\n* 新しいエンティティの作成\n* 既存のエンティティの更新\n* さらに、特定のMerge API設定に基づいて\n\n\\*\\*注:\\*\\*ダウンロードツールは現在サポートされていません。これは既知の制限であり、今後のリリースで修正される予定です。\n\n## 🔑 環境変数\n\nMerge MCP サーバーでは次の環境変数が使用されます。\n\n* `MERGE_API_KEY` : Merge APIキー\n* `MERGE_ACCOUNT_TOKEN` : マージリンクアカウントトークン\n* `MERGE_TENANT` (オプション): Merge APIテナント。有効な値は`US` 、 `EU` 、 `APAC`です。デフォルトは`US`です。","Japanese","ja-JP",{"_36":62,"_40":66,"_45":67},{"_38":63,"_40":64,"_42":65},"MCP 프로토콜을 지원하는 Merge API와 LLM 공급자 간의 통합을 제공하여 HRIS, ATS 및 기타 범주에서 Merge 데이터와 자연어 상호 작용을 허용합니다.","MCP 서버 병합","# MCP 서버 병합\n\n이 MCP(모델 컨텍스트 프로토콜) 서버는 Merge API와 MCP 프로토콜을 지원하는 모든 LLM 공급자(예: 데스크톱용 Claude) 간의 통합을 제공하여 자연어를 사용하여 Merge 데이터와 상호 작용할 수 있도록 합니다.\n\n## ✨ 특징\n\n* 자연어를 사용하여 API 엔터티 병합 쿼리\n* 병합 데이터 모델과 해당 필드에 대한 정보를 얻으세요.\n* 대화형 인터페이스를 통해 엔터티를 생성하고 업데이트합니다.\n* 다양한 Merge API 카테고리(HRIS, ATS 등) 지원\n\n## 📦 설치\n\n### 필수 조건\n\n* API 키와 계정 토큰 병합\n* Python 3.10 이상\n* [자외선](https://github.com/astral-sh/uv)\n\n독립 실행형 설치 프로그램으로 `uv` 설치하세요:\n\n지엑스피1\n\n또는 pip를 통해:\n\n```bash\n# With pip.\npip install uv\n\n# With pipx.\npipx install uv\n```\n\n## 🔌 MCP 설정\n\nMerge MCP를 설정하는 데 사용할 수 있는 구성 파일의 예는 다음과 같습니다.\n\n```json\n{\n \"mcpServers\": {\n \"merge-mcp-server\": {\n \"command\": \"uvx\",\n \"args\": [\"merge-mcp\"],\n \"env\": {\n \"MERGE_API_KEY\": \"your_api_key\",\n \"MERGE_ACCOUNT_TOKEN\": \"your_account_token\"\n }\n }\n }\n}\n```\n\n참고: \"uvx\" 명령이 작동하지 않으면 절대 경로(예: /Users/username/.local/bin/uvx)를 시도해 보세요.\n\n### Claude Desktop 구성 예시\n\n1. `uvx` 설치되어 있는지 확인하세요\n\n2. 공식 웹사이트에서 [Claude Desktop을](https://claude.ai/download) 다운로드하세요\n\n3. 다운로드가 완료되면 앱을 열고 지침에 따라 계정을 설정하세요.\n\n4. **설정 → 개발자 → 구성 편집** 으로 이동하세요. 그러면 `claude_desktop_config.json` 파일이 텍스트 편집기에 열립니다.\n\n5. 위의 MCP 서버 설정 JSON을 복사하여 텍스트 편집기에 붙여넣습니다.\n\n6. `your_api_key` 와 `your_account_token` 실제 Merge API 키와 연결된 계정 토큰으로 바꾸세요. 또한 `uvx` 설정 파일에 있는 명령의 절대 경로(예: `/Users/username/.local/bin/uvx` )로 바꿔야 합니다. 터미널에서 `which uvx` 명령을 실행하면 절대 경로를 확인할 수 있습니다.\n\n7. 구성 파일을 저장합니다\n\n8. 도구를 보려면 Claude Desktop을 다시 시작하세요. 도구가 나타나기까지 1분 정도 걸릴 수 있습니다.\n\n### Python 클라이언트 구성 예\n\n1. 환경 설정\n\n```bash\n# Create project directory\nmkdir mcp-client\ncd mcp-client\n\n# Create virtual environment\npython -m venv .venv\n\n# Activate virtual environment\n# On Windows:\n.venv\\Scripts\\activate\n# On Unix or MacOS:\nsource .venv/bin/activate\n\n# Install required packages\npip install mcp uv anthropic python-dotenv\n\n# Create our main file\ntouch client.py\n```\n\n2. API 키 설정\n\n```bash\n# Add your ANTHROPIC_API_KEY and MERGE_API_KEY to .env\necho \"ANTHROPIC_API_KEY=\" >> .env\necho \"MERGE_API_KEY=\" >> .env\n\n# Add .env file to .gitignore\necho \".env\" >> .gitignore\n```\n\n3. `client.py` 파일을 만들고 다음 코드를 추가합니다.\n\n```python\nimport os\nimport asyncio\nfrom typing import Optional\nfrom contextlib import AsyncExitStack\n\nfrom mcp import ClientSession, StdioServerParameters\nfrom mcp.client.stdio import stdio_client\n\nfrom anthropic import Anthropic\nfrom dotenv import load_dotenv\n\nload_dotenv() # load environment variables from .env\n\nclass MCPClient:\n def __init__(self):\n # Initialize session and client objects\n self.session: Optional[ClientSession] = None\n self.exit_stack = AsyncExitStack()\n self.anthropic = Anthropic()\n\n # Methods will go here\n```\n\n4. MCPClient 클래스에 `connect_to_server` 함수를 추가합니다.\n\n```python\nasync def connect_to_server(self, linked_account_token: str):\n \"\"\"Connect to an MCP server\n Args:\n linked_account_token: The token for the associated Linked Account\n \"\"\"\n\n server_params = StdioServerParameters(\n command=\"uvx\",\n args=[\"merge-mcp\"],\n env={\n \"MERGE_API_KEY\": os.getenv(\"MERGE_API_KEY\"),\n \"MERGE_ACCOUNT_TOKEN\": linked_account_token\n }\n )\n\n stdio_transport = await self.exit_stack.enter_async_context(stdio_client(server_params))\n self.stdio, self.write = stdio_transport\n self.session = await self.exit_stack.enter_async_context(ClientSession(self.stdio, self.write))\n\n await self.session.initialize()\n\n # List available tools\n response = await self.session.list_tools()\n tools = response.tools\n print(\"\\nConnected to server with tools:\", [tool.name for tool in tools])\n```\n\n5. MCPClient 클래스에 `process_query` 함수를 추가합니다.\n\n```python\nasync def process_query(self, query: str) -> str:\n \"\"\"Process a query using Claude and available tools\"\"\"\n messages = [\n {\n \"role\": \"user\",\n \"content\": query\n }\n ]\n\n response = await self.session.list_tools()\n available_tools = [{\n \"name\": tool.name,\n \"description\": tool.description,\n \"input_schema\": tool.inputSchema\n } for tool in response.tools]\n\n # Initial Claude API call\n response = self.anthropic.messages.create(\n model=\"claude-3-5-sonnet-20241022\",\n max_tokens=1000,\n messages=messages,\n tools=available_tools\n )\n\n # Process response and handle tool calls\n final_text = []\n assistant_message_content = []\n for content in response.content:\n if content.type == 'text':\n final_text.append(content.text)\n assistant_message_content.append(content)\n\n elif content.type == 'tool_use':\n tool_name = content.name\n tool_args = content.input\n\n # Get confirmation for tool call execution\n confirmation = input(f\"Do you want to call tool '{tool_name}' with arguments {tool_args}? (y/n): \").strip().lower()\n if confirmation.startswith('y'):\n result = await self.session.call_tool(tool_name, tool_args)\n final_text.append(f\"[Calling tool {tool_name} with args {tool_args}]\")\n assistant_message_content.append(content)\n messages.append({\n \"role\": \"assistant\",\n \"content\": assistant_message_content\n })\n messages.append({\n \"role\": \"user\",\n \"content\": [\n {\n \"type\": \"tool_result\",\n \"tool_use_id\": content.id,\n \"content\": result.content\n }\n ]\n })\n\n # Get next response from Claude\n response = self.anthropic.messages.create(\n model=\"claude-3-5-sonnet-20241022\",\n max_tokens=1000,\n messages=messages,\n tools=available_tools\n )\n final_text.append(response.content[0].text)\n\n else:\n final_text.append(f\"[Skipped calling tool {tool_name} with args {tool_args}]\")\n\n return \"\\n\".join(final_text)\n```\n\n6. MCPClient 클래스에 `chat_loop` 함수를 추가합니다.\n\n```python\n async def chat_loop(self):\n \"\"\"Run an interactive chat loop\"\"\"\n print(\"\\nMCP Client Started!\")\n print(\"Type your queries or 'quit' to exit.\")\n\n while True:\n try:\n query = input(\"\\nQuery: \").strip()\n\n if query.lower() == 'quit':\n break\n\n response = await self.process_query(query)\n print(\"\\n\" + response)\n\n except Exception as e:\n print(f\"\\nError: {str(e)}\")\n```\n\n7. MCPClient 클래스에 `cleanup` 기능 추가\n\n```python\n async def cleanup(self):\n \"\"\"Clean up resources\"\"\"\n await self.exit_stack.aclose()\n```\n\n8. `client.py` 파일에 주요 진입점으로 `main` 함수를 추가합니다.\n\n```python\nasync def main():\n client = MCPClient()\n try:\n await client.connect_to_server(\"\")\n await client.chat_loop()\n finally:\n await client.cleanup()\n\nif __name__ == \"__main__\":\n import sys\n asyncio.run(main())\n```\n\n9. 클라이언트 실행\n\n```bash\npython client.py\n```\n\n## 🔍 범위\n\n범위는 MCP 서버에서 활성화되는 도구를 결정하며, Merge API의 여러 부분에 대한 액세스를 제어하는 데 사용됩니다. **범위를 지정하지 않으면 사용 가능한 모든 범위가 활성화됩니다.**\n\n서버를 시작할 때 활성화할 범위를 지정할 수 있습니다. 이는 `--scopes` 플래그와 범위 목록을 전달하여 수행됩니다.\n\n```json\n{\n \"mcpServers\": {\n \"merge-mcp-server\": {\n \"command\": \"uvx\",\n \"args\": [\n \"merge-mcp\",\n \"--scopes\",\n \"ats.Job:read\",\n \"ats.Candidate\",\n \"ats.Application:write\"\n ],\n \"env\": {\n \"MERGE_API_KEY\": \"your_api_key\",\n \"MERGE_ACCOUNT_TOKEN\": \"your_account_token\"\n }\n }\n }\n}\n```\n\n### 범위 형식\n\nMerge MCP 서버의 범위는 Merge API 범주와 일반 모델 이름에 따라 특정 형식을 따릅니다. 각 범위는 다음과 같은 형식을 갖습니다.\n\n```\n.:\n```\n\n어디:\n\n* `` 는 Merge API 카테고리(예: `hris` , `ats` , `accounting` )입니다.\n* `` 은 병합 공통 모델의 이름입니다(예: `Employee` , `Candidate` , `Account` ).\n* `` 은 `read` 또는 `write` 입니다(선택 사항 - 지정하지 않으면 모든 권한이 부여됨)\n\n유효한 범위의 예:\n\n* `hris.Employee:read` - HRIS 카테고리에서 직원 데이터를 읽을 수 있습니다.\n* `ats.Candidate:write` - ATS 카테고리에서 후보자 데이터를 생성하거나 업데이트할 수 있습니다.\n* `accounting.Account` - 회계 범주의 계정 데이터에 대한 모든 작업을 허용합니다.\n\n여러 범위를 결합하여 다양한 권한을 부여할 수 있습니다.\n\n### 범위 가용성에 대한 중요 참고 사항\n\n사용 가능한 범위는 Merge API 계정 구성과 연결된 계정에서 액세스할 수 있는 모델에 따라 달라집니다. 범위는 연결된 계정에서 활성화된 범위와 상호 참조되어야 합니다.\n\n* **범주 불일치** : 연결된 계정과 일치하지 않는 범주에 대한 범위를 지정하는 경우(예: HRIS 연결된 계정과 함께 `ats.Job` 사용), 해당 범위에 대한 도구는 반환되지 않습니다.\n\n* **권한 불일치** : 연결된 계정에 대해 활성화되지 않은 권한을 요청하는 경우(예: 읽기 액세스만 활성화된 상태에서 `hris.Employee:write` 사용), 해당 권한이 필요한 도구는 반환되지 않습니다.\n\n* **검증** : 서버는 귀하가 요청한 범위를 귀하의 연결된 계정에서 사용 가능한 범위와 자동으로 비교하고, 유효하고 승인된 범위에 대한 도구만 활성화합니다.\n\n범위는 일반적으로 Merge API의 다양한 모델이나 엔터티 유형에 해당하며, 이러한 엔터티에 대한 읽기 및 쓰기 액세스를 모두 제어합니다.\n\n## 🚀 사용 가능한 도구\n\nMerge MCP 서버는 다양한 Merge API 엔드포인트에 대한 액세스를 도구로 제공합니다. 사용 가능한 도구는 Merge API 범주(HRIS, ATS 등)와 활성화된 범위에 따라 달라집니다.\n\n도구는 Merge API 스키마를 기반으로 동적으로 생성되며 다음 작업을 포함합니다.\n\n* 엔터티 세부 정보 검색\n* 엔터티 목록\n* 새로운 엔터티 생성\n* 기존 엔터티 업데이트\n* 그리고 귀하의 특정 Merge API 구성에 따라 더 많은 기능이 제공됩니다.\n\n**참고:** 다운로드 도구는 현재 지원되지 않습니다. 이는 알려진 제한 사항이며 향후 릴리스에서 해결될 예정입니다.\n\n## 🔑 환경 변수\n\nMerge MCP 서버에서는 다음 환경 변수를 사용합니다.\n\n* `MERGE_API_KEY` : Merge API 키\n* `MERGE_ACCOUNT_TOKEN` : Merge 연결 계정 토큰\n* `MERGE_TENANT` (선택 사항): Merge API 테넌트입니다. 유효한 값은 `US` , `EU` , `APAC` 입니다. 기본값은 `US` 입니다.","Korean","ko-KR",{"_36":69,"_40":73,"_45":74},{"_38":70,"_40":71,"_42":72},"提供 Merge API 与支持 MCP 协议的 LLM 提供商之间的集成，允许跨 HRIS、ATS 和其他类别与 Merge 数据进行自然语言交互。","合并 MCP 服务器","# 合并 MCP 服务器\n\n该 MCP（模型上下文协议）服务器提供 Merge API 与任何支持 MCP 协议的 LLM 提供程序（例如，Claude for Desktop）之间的集成，允许您使用自然语言与 Merge 数据进行交互。\n\n## ✨ 特点\n\n* 使用自然语言查询合并 API 实体\n* 获取有关合并数据模型及其字段的信息\n* 通过对话界面创建和更新实体\n* 支持多种合并 API 类别（HRIS、ATS 等）\n\n## 📦安装\n\n### 先决条件\n\n* 合并 API 密钥和帐户令牌\n* Python 3.10 或更高版本\n* [紫外线](https://github.com/astral-sh/uv)\n\n使用独立安装程序安装`uv` ：\n\n```bash\n# On macOS and Linux.\ncurl -LsSf https://astral.sh/uv/install.sh | sh\n\n# On Windows.\npowershell -ExecutionPolicy ByPass -c \"irm https://astral.sh/uv/install.ps1 | iex\"\n```\n\n或者通过 pip：\n\n```bash\n# With pip.\npip install uv\n\n# With pipx.\npipx install uv\n```\n\n## 🔌 MCP 设置\n\n这是一个示例配置文件，您可以用它来设置合并 MCP。\n\n```json\n{\n \"mcpServers\": {\n \"merge-mcp-server\": {\n \"command\": \"uvx\",\n \"args\": [\"merge-mcp\"],\n \"env\": {\n \"MERGE_API_KEY\": \"your_api_key\",\n \"MERGE_ACCOUNT_TOKEN\": \"your_account_token\"\n }\n }\n }\n}\n```\n\n注意：如果“uvx”命令不起作用，请尝试绝对路径（即/Users/username/.local/bin/uvx）\n\n### Claude Desktop 配置示例\n\n1. 确保你已经安装了`uvx`\n\n2. 从官方网站下载[Claude Desktop](https://claude.ai/download)\n\n3. 下载后，打开应用程序并按照说明设置您的帐户\n\n4. 前往\\*\\*“设置”→“开发者”→“编辑配置”\\*\\* 。这将在文本编辑器中打开一个名为`claude_desktop_config.json`的文件。\n\n5. 复制上面的 MCP 服务器设置 JSON 并将其粘贴到文本编辑器中\n\n6. 将`your_api_key`和`your_account_token`替换为你实际的合并 API 密钥和关联账户令牌。你还需要将`uvx`替换为配置文件中命令的绝对路径（例如`/Users/username/.local/bin/uvx` ）。你可以在终端中运行`which uvx`来获取绝对路径。\n\n7. 保存配置文件\n\n8. 重启 Claude Desktop 即可查看您的工具。工具可能需要一分钟才能显示。\n\n### Python 客户端配置示例\n\n1. 设置您的环境\n\n```bash\n# Create project directory\nmkdir mcp-client\ncd mcp-client\n\n# Create virtual environment\npython -m venv .venv\n\n# Activate virtual environment\n# On Windows:\n.venv\\Scripts\\activate\n# On Unix or MacOS:\nsource .venv/bin/activate\n\n# Install required packages\npip install mcp uv anthropic python-dotenv\n\n# Create our main file\ntouch client.py\n```\n\n2. 设置您的 API 密钥\n\n```bash\n# Add your ANTHROPIC_API_KEY and MERGE_API_KEY to .env\necho \"ANTHROPIC_API_KEY=\" >> .env\necho \"MERGE_API_KEY=\" >> .env\n\n# Add .env file to .gitignore\necho \".env\" >> .gitignore\n```\n\n3. 创建`client.py`文件，添加以下代码\n\n```python\nimport os\nimport asyncio\nfrom typing import Optional\nfrom contextlib import AsyncExitStack\n\nfrom mcp import ClientSession, StdioServerParameters\nfrom mcp.client.stdio import stdio_client\n\nfrom anthropic import Anthropic\nfrom dotenv import load_dotenv\n\nload_dotenv() # load environment variables from .env\n\nclass MCPClient:\n def __init__(self):\n # Initialize session and client objects\n self.session: Optional[ClientSession] = None\n self.exit_stack = AsyncExitStack()\n self.anthropic = Anthropic()\n\n # Methods will go here\n```\n\n4. 向 MCPClient 类添加`connect_to_server`函数\n\n```python\nasync def connect_to_server(self, linked_account_token: str):\n \"\"\"Connect to an MCP server\n Args:\n linked_account_token: The token for the associated Linked Account\n \"\"\"\n\n server_params = StdioServerParameters(\n command=\"uvx\",\n args=[\"merge-mcp\"],\n env={\n \"MERGE_API_KEY\": os.getenv(\"MERGE_API_KEY\"),\n \"MERGE_ACCOUNT_TOKEN\": linked_account_token\n }\n )\n\n stdio_transport = await self.exit_stack.enter_async_context(stdio_client(server_params))\n self.stdio, self.write = stdio_transport\n self.session = await self.exit_stack.enter_async_context(ClientSession(self.stdio, self.write))\n\n await self.session.initialize()\n\n # List available tools\n response = await self.session.list_tools()\n tools = response.tools\n print(\"\\nConnected to server with tools:\", [tool.name for tool in tools])\n```\n\n5. 向 MCPClient 类添加`process_query`函数\n\n```python\nasync def process_query(self, query: str) -> str:\n \"\"\"Process a query using Claude and available tools\"\"\"\n messages = [\n {\n \"role\": \"user\",\n \"content\": query\n }\n ]\n\n response = await self.session.list_tools()\n available_tools = [{\n \"name\": tool.name,\n \"description\": tool.description,\n \"input_schema\": tool.inputSchema\n } for tool in response.tools]\n\n # Initial Claude API call\n response = self.anthropic.messages.create(\n model=\"claude-3-5-sonnet-20241022\",\n max_tokens=1000,\n messages=messages,\n tools=available_tools\n )\n\n # Process response and handle tool calls\n final_text = []\n assistant_message_content = []\n for content in response.content:\n if content.type == 'text':\n final_text.append(content.text)\n assistant_message_content.append(content)\n\n elif content.type == 'tool_use':\n tool_name = content.name\n tool_args = content.input\n\n # Get confirmation for tool call execution\n confirmation = input(f\"Do you want to call tool '{tool_name}' with arguments {tool_args}? (y/n): \").strip().lower()\n if confirmation.startswith('y'):\n result = await self.session.call_tool(tool_name, tool_args)\n final_text.append(f\"[Calling tool {tool_name} with args {tool_args}]\")\n assistant_message_content.append(content)\n messages.append({\n \"role\": \"assistant\",\n \"content\": assistant_message_content\n })\n messages.append({\n \"role\": \"user\",\n \"content\": [\n {\n \"type\": \"tool_result\",\n \"tool_use_id\": content.id,\n \"content\": result.content\n }\n ]\n })\n\n # Get next response from Claude\n response = self.anthropic.messages.create(\n model=\"claude-3-5-sonnet-20241022\",\n max_tokens=1000,\n messages=messages,\n tools=available_tools\n )\n final_text.append(response.content[0].text)\n\n else:\n final_text.append(f\"[Skipped calling tool {tool_name} with args {tool_args}]\")\n\n return \"\\n\".join(final_text)\n```\n\n6. 向 MCPClient 类添加`chat_loop`函数\n\n```python\n async def chat_loop(self):\n \"\"\"Run an interactive chat loop\"\"\"\n print(\"\\nMCP Client Started!\")\n print(\"Type your queries or 'quit' to exit.\")\n\n while True:\n try:\n query = input(\"\\nQuery: \").strip()\n\n if query.lower() == 'quit':\n break\n\n response = await self.process_query(query)\n print(\"\\n\" + response)\n\n except Exception as e:\n print(f\"\\nError: {str(e)}\")\n```\n\n7. 向 MCPClient 类添加`cleanup`函数\n\n```python\n async def cleanup(self):\n \"\"\"Clean up resources\"\"\"\n await self.exit_stack.aclose()\n```\n\n8. 在`client.py`文件中添加一个`main`函数作为主入口点\n\n```python\nasync def main():\n client = MCPClient()\n try:\n await client.connect_to_server(\"\")\n await client.chat_loop()\n finally:\n await client.cleanup()\n\nif __name__ == \"__main__\":\n import sys\n asyncio.run(main())\n```\n\n9. 运行客户端\n\n```bash\npython client.py\n```\n\n## 🔍 范围\n\n范围决定了 MCP 服务器上启用哪些工具，并用于控制对 Merge API 各个部分的访问。**如果未指定范围，则所有可用的范围均会启用。**\n\n启动服务器时，您可以指定要启用的范围。这可以通过传递带有范围列表的`--scopes`标志来实现。\n\n```json\n{\n \"mcpServers\": {\n \"merge-mcp-server\": {\n \"command\": \"uvx\",\n \"args\": [\n \"merge-mcp\",\n \"--scopes\",\n \"ats.Job:read\",\n \"ats.Candidate\",\n \"ats.Application:write\"\n ],\n \"env\": {\n \"MERGE_API_KEY\": \"your_api_key\",\n \"MERGE_ACCOUNT_TOKEN\": \"your_account_token\"\n }\n }\n }\n}\n```\n\n### 范围格式\n\nMerge MCP 服务器中的范围遵循基于 Merge API 类别和通用模型名称的特定格式。每个范围的格式如下：\n\n```\n.:\n```\n\n在哪里：\n\n* ``是合并 API 类别（例如`hris` 、 `ats` 、 `accounting` ）\n* ``是合并通用模型的名称（例如， `Employee` 、 `Candidate` 、 `Account` ）\n* ``是`read`或`write` （可选 - 如果未指定，则授予所有权限）\n\n有效范围的示例：\n\n* `hris.Employee:read` - 允许从 HRIS 类别读取员工数据\n* `ats.Candidate:write` - 允许在 ATS 类别中创建或更新候选人数据\n* `accounting.Account` - 允许对会计类别中的帐户数据进行所有操作\n\n您可以组合多个范围来授予不同的权限。\n\n### 关于范围可用性的重要说明\n\n可用的范围取决于您的合并 API 帐户配置以及关联帐户可以访问的模型。范围必须与关联帐户中已启用的范围进行交叉引用：\n\n* **类别不匹配**：如果您指定的类别范围与您的关联帐户不匹配（例如，使用带有 HRIS 关联帐户的`ats.Job` ），则不会返回该范围的任何工具。\n\n* **权限不匹配**：如果您请求的权限未为您的链接帐户启用（例如，在仅启用读取访问权限时使用`hris.Employee:write` ），则不会返回需要该权限的工具。\n\n* **验证**：服务器将自动验证您请求的范围是否与您的链接帐户中可用的范围一致，并且仅为有效的、授权的范围启用工具。\n\n范围通常对应于 Merge API 中的不同模型或实体类型，它们控制对这些实体的读取和写入访问。\n\n## 🚀 可用工具\n\n合并 MCP 服务器提供对各种合并 API 端点的访问，这些端点以工具形式提供。可用的工具取决于您的合并 API 类别（HRIS、ATS 等）以及您启用的范围。\n\n工具是根据您的合并 API 模式动态生成的，并包括以下操作：\n\n* 检索实体详细信息\n* 列出实体\n* 创建新实体\n* 更新现有实体\n* 更多内容取决于您的具体合并 API 配置\n\n\\*\\*注意：\\*\\*目前不支持下载工具。这是一个已知限制，将在后续版本中解决。\n\n## 🔑 环境变量\n\nMerge MCP 服务器使用以下环境变量：\n\n* `MERGE_API_KEY` ：您的合并 API 密钥\n* `MERGE_ACCOUNT_TOKEN` ：您的合并关联账户令牌\n* `MERGE_TENANT` （可选）：合并 API 租户。有效值为`US` 、 `EU`和`APAC` 。默认为`US` 。","Chinese","zh-CN","githubUser","locale",{"_78":79,"_80":81,"_20":99,"_108":39,"_109":39,"_110":111,"_114":115,"_135":-5,"_23":136,"_137":138,"_139":140,"_40":41,"_141":128,"_142":-5,"_143":-5,"_144":-5,"_145":-5,"_146":147,"_151":-5,"_152":124,"_153":154,"_158":-5,"_8":159,"_160":161},"addedAt","2025-04-08T21:02:51.615615983Z","argumentsJsonSchema",{"_82":83,"_96":97,"_12":98},"properties",{"_84":85,"_90":91,"_93":94},"MERGE_TENANT",{"_86":87,"_38":88,"_12":89},"default","US","The Merge API tenant (US, EU, APAC). Defaults to US","string","MERGE_API_KEY",{"_38":92,"_12":89},"Your Merge API key","MERGE_ACCOUNT_TOKEN",{"_38":95,"_12":89},"Your Merge Linked Account token","required",[90,93],"object",[100,101,102,103,104,105,106,107],"author:official","category:rag-systems","category:databases","category:api-testing","environment:network","hosting:remote-capable","language:python","license_grade:f","descriptionMarkdown","descriptionPlainText","githubProject",{"_42":43,"_112":113},"url","https://github.com/merge-api/merge-mcp","githubRepository",{"_116":117,"_118":119,"_120":121,"_122":123,"_40":124,"_125":126,"_130":131,"_132":-5,"_133":134},"defaultBranch","main","fullName","merge-api/merge-mcp","githubTreeId",113383,"language","Python","merge-mcp","owner",{"_127":128,"_8":129},"login","merge-api","r3ySDGh0CJ","recentStargazers",0,"spdxLicense","stargazers",7,"glamaJson",5210,"inspectable",false,"integrations",[],"namespace","npmPackage","prompts","redditPostForeignId","resources","scores",{"_148":131,"_149":-5,"_150":-5},"license","quality","security","simpleIcon","slug","supportedPlatforms",[155,156,157],"MACOS","WINDOWS","LINUX","tools","1qfe7rig9s","updatedAt","2025-05-01T05:00:32.762762711Z","meta",[164,166,168,175,178,180,183,186,189,192,195,197,200,203,205,207,209,211,213],{"_10":165},"Merge MCP Server | Glama",{"_167":39,"_40":38},"content",{"_169":170,"_171":172,"_173":174},"href","https://glama.ai/mcp/servers/@merge-api/merge-mcp","rel","canonical","tagName","link",{"_167":41,"_176":177},"property","og:title",{"_167":39,"_176":179},"og:description",{"_167":181,"_176":182},"https://glama.ai/logo.png","og:logo",{"_167":184,"_176":185},"https://glama.ai/mcp/servers/@merge-api/merge-mcp/og-image","og:image",{"_167":187,"_176":188},"630","og:image:height",{"_167":190,"_176":191},"1200","og:image:width",{"_167":193,"_176":194},"website","og:type",{"_167":170,"_176":196},"og:url",{"_167":198,"_176":199},"Glama – MCP Hosting Platform","og:site_name",{"_167":201,"_40":202},"summary_large_image","twitter:card",{"_167":198,"_40":204},"twitter:title",{"_167":39,"_40":206},"twitter:description",{"_167":184,"_40":208},"twitter:image",{"_167":187,"_40":210},"twitter:image:height",{"_167":190,"_40":212},"twitter:image:width",{"_214":215},"script:ld+json",{"_216":217,"_218":219},"@context","http://schema.org","@graph",[220],{"_216":221,"_222":223,"_224":225},"https://schema.org","@type","BreadcrumbList","itemListElement",[226,233,237,241],{"_222":227,"_228":229,"_40":230,"_231":232},"ListItem","item","https://glama.ai/","Glama","position",1,{"_222":227,"_228":234,"_40":235,"_231":236},"https://glama.ai/mcp","MCP",2,{"_222":227,"_228":238,"_40":239,"_231":240},"https://glama.ai/mcp/servers","Servers",3,{"_222":227,"_228":170,"_40":41,"_231":242},4,"visitorIsMaintainer","routes/_public/mcp/servers/~namespace/~slug/_pages/reviews/_route",{"_75":-5,"_36":246,"_162":260,"_295":-5},{"_78":79,"_80":247,"_20":253,"_108":39,"_109":39,"_110":254,"_114":255,"_135":-5,"_23":136,"_137":138,"_139":257,"_40":41,"_141":128,"_142":-5,"_143":-5,"_144":-5,"_145":-5,"_146":258,"_151":-5,"_152":124,"_153":259,"_158":-5,"_8":159,"_160":161},{"_82":248,"_96":252,"_12":98},{"_84":249,"_90":250,"_93":251},{"_86":87,"_38":88,"_12":89},{"_38":92,"_12":89},{"_38":95,"_12":89},[90,93],[100,101,102,103,104,105,106,107],{"_42":43,"_112":113},{"_116":117,"_118":119,"_120":121,"_122":123,"_40":124,"_125":256,"_130":131,"_132":-5,"_133":134},{"_127":128,"_8":129},[],{"_148":131,"_149":-5,"_150":-5},[155,156,157],[261,263,265,267,269,270,271,272,273,274,275,276,277,278,279,280,281,282,283],{"_10":262},"Reviews | Merge MCP Server | Glama",{"_167":264,"_40":38},"Read community reviews and feedback for Merge MCP Server. Discover user experiences, ratings, strengths, and potential limitations of this MCP implementation from actual users.",{"_169":266,"_171":172,"_173":174},"https://glama.ai/mcp/servers/@merge-api/merge-mcp/reviews",{"_167":268,"_176":177},"Reviews | Merge MCP Server",{"_167":264,"_176":179},{"_167":181,"_176":182},{"_167":184,"_176":185},{"_167":187,"_176":188},{"_167":190,"_176":191},{"_167":193,"_176":194},{"_167":266,"_176":196},{"_167":198,"_176":199},{"_167":201,"_40":202},{"_167":198,"_40":204},{"_167":264,"_40":206},{"_167":184,"_40":208},{"_167":187,"_40":210},{"_167":190,"_40":212},{"_214":284},{"_216":217,"_218":285},[286],{"_216":221,"_222":223,"_224":287},[288,289,290,291,292],{"_222":227,"_228":229,"_40":230,"_231":232},{"_222":227,"_228":234,"_40":235,"_231":236},{"_222":227,"_228":238,"_40":239,"_231":240},{"_222":227,"_228":170,"_40":41,"_231":242},{"_222":227,"_228":266,"_40":293,"_231":294},"Reviews",5,"review","actionData","errors"]

Rate & Review

MCP directory API