Un complemento de Protocolo de contexto de modelo (MCP) que proporciona acceso perfecto a las potentes capacidades de Grok AI directamente desde Cline.
Este complemento expone tres herramientas poderosas a través de la interfaz MCP:
Completar chat : genere respuestas de texto utilizando los modelos de lenguaje de Grok
Comprensión de imágenes : analice imágenes con las capacidades de visión de Grok
Llamada de funciones : utilice Grok para llamar a funciones según la entrada del usuario
Related MCP server: Grok MCP Server
Node.js (v16 o superior)
Una clave API de Grok AI (obtenida en console.x.ai )
Cline con soporte MCP
Clonar este repositorio:
git clone https://github.com/Bob-lance/grok-mcp.git
cd grok-mcpInstalar dependencias:
npm installConstruir el proyecto:
npm run buildAgregue el servidor MCP a su configuración de Cline MCP:
Para la extensión Cline de VSCode, edite el archivo en:
~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.jsonAgregue la siguiente configuración:
{
"mcpServers": {
"grok-mcp": {
"command": "node",
"args": ["/path/to/grok-mcp/build/index.js"],
"env": {
"XAI_API_KEY": "your-grok-api-key"
},
"disabled": false,
"autoApprove": []
}
}
}Reemplace /path/to/grok-mcp con la ruta real a su instalación y your-grok-api-key con su clave API de Grok AI.
Una vez instalado y configurado, el complemento Grok MCP proporciona tres herramientas que se pueden utilizar en Cline:
Genere respuestas de texto utilizando los modelos de lenguaje de Grok:
<use_mcp_tool>
<server_name>grok-mcp</server_name>
<tool_name>chat_completion</tool_name>
<arguments>
{
"messages": [
{
"role": "system",
"content": "You are a helpful assistant."
},
{
"role": "user",
"content": "Hello, what can you tell me about Grok AI?"
}
],
"temperature": 0.7
}
</arguments>
</use_mcp_tool>Analice imágenes con las capacidades de visión de Grok:
<use_mcp_tool>
<server_name>grok-mcp</server_name>
<tool_name>image_understanding</tool_name>
<arguments>
{
"image_url": "https://example.com/image.jpg",
"prompt": "What is shown in this image?"
}
</arguments>
</use_mcp_tool>También puedes utilizar imágenes codificadas en base64:
<use_mcp_tool>
<server_name>grok-mcp</server_name>
<tool_name>image_understanding</tool_name>
<arguments>
{
"base64_image": "base64-encoded-image-data",
"prompt": "What is shown in this image?"
}
</arguments>
</use_mcp_tool>Utilice Grok para llamar funciones según la entrada del usuario:
<use_mcp_tool>
<server_name>grok-mcp</server_name>
<tool_name>function_calling</tool_name>
<arguments>
{
"messages": [
{
"role": "user",
"content": "What's the weather like in San Francisco?"
}
],
"tools": [
{
"type": "function",
"function": {
"name": "get_weather",
"description": "Get the current weather in a given location",
"parameters": {
"type": "object",
"properties": {
"location": {
"type": "string",
"description": "The city and state, e.g. San Francisco, CA"
},
"unit": {
"type": "string",
"enum": ["celsius", "fahrenheit"],
"description": "The unit of temperature to use"
}
},
"required": ["location"]
}
}
}
]
}
</arguments>
</use_mcp_tool>Genere una respuesta utilizando la función de finalización de chat de Grok AI.
Parámetros:
messages (obligatorio): Matriz de objetos de mensaje con rol y contenido
model (opcional): modelo de Grok a utilizar (predeterminado: grok-2-latest)
temperature (opcional): Temperatura de muestreo (0-2, predeterminado 1)
max_tokens (opcional): Número máximo de tokens a generar (predeterminado: 16384)
Analice imágenes utilizando las capacidades de visión de Grok AI.
Parámetros:
prompt (obligatorio): Texto de aviso para acompañar la imagen
image_url (opcional): URL de la imagen a analizar
base64_image (opcional): datos de imagen codificados en Base64 (sin el prefijo data:image)
model (opcional): modelo de visión de Grok a utilizar (el valor predeterminado es grok-2-vision-latest)
Nota: Se debe proporcionar image_url o base64_image .
Utilice Grok AI para llamar funciones según la entrada del usuario.
Parámetros:
messages (obligatorio): Matriz de objetos de mensaje con rol y contenido
tools (obligatorio): Matriz de objetos de herramientas con tipo, nombre de función, descripción y parámetros
tool_choice (opcional): modo de elección de herramienta (automático, obligatorio, ninguno, predeterminado en automático)
model (opcional): modelo de Grok a utilizar (predeterminado: grok-2-latest)
src/index.ts - Implementación del servidor principal
src/grok-api-client.ts - Implementación del cliente de la API de Grok
npm run buildXAI_API_KEY="your-grok-api-key" node build/index.jsEste proyecto está licenciado bajo la licencia MIT: consulte el archivo de LICENCIA para obtener más detalles.
Generate a response using Grok AI chat completion
| Name | Required | Description | Default |
|---|---|---|---|
| max_tokens | No | Maximum number of tokens to generate | |
| messages | Yes | Array of message objects with role and content | |
| model | No | Grok model to use (e.g., grok-2-latest, grok-3, grok-3-reasoner, grok-3-deepsearch, grok-3-mini-beta) | grok-3-mini-beta |
| temperature | No | Sampling temperature (0-2) |
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 but offers minimal information. It states what the tool does but doesn't describe rate limits, authentication requirements, response formats, error conditions, or any operational constraints. For a generative AI tool with significant behavioral implications, this is inadequate.
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 states the core purpose without unnecessary elaboration. It's appropriately sized for a tool with comprehensive schema documentation and gets straight to the point with zero 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 generative AI tool with no annotations and no output schema, the description is insufficient. It doesn't explain what kind of response is generated, how to interpret results, error handling, or operational constraints. The agent lacks crucial context about this tool's behavior and outputs despite the comprehensive input schema.
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%, so the schema fully documents all 4 parameters. The description adds no parameter-specific information beyond what's already in the schema. According to scoring rules, when schema coverage is high (>80%), the baseline is 3 even with no param info in 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 clearly states the action ('Generate a response') and the resource/technology ('using Grok AI chat completion'), which is specific and unambiguous. However, it doesn't differentiate this tool from its sibling tools (function_calling, image_understanding) - all three appear to be different Grok AI capabilities, but the description doesn't explain how chat completion differs from function calling or image understanding.
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 its siblings. There's no mention of appropriate contexts for chat completion versus function calling or image understanding, nor any prerequisites or constraints. The agent must infer usage from the tool name alone.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Use Grok AI to call functions based on user input
| Name | Required | Description | Default |
|---|---|---|---|
| messages | Yes | Array of message objects with role and content | |
| model | No | Grok model to use (e.g., grok-2-latest, grok-3, grok-3-reasoner, grok-3-deepsearch, grok-3-mini-beta) | grok-3-mini-beta |
| tool_choice | No | Tool choice mode (auto, required, none) | auto |
| tools | Yes | Array of tool objects with type, function name, description, and parameters |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries full burden but only states the basic action without disclosing behavioral traits like rate limits, authentication needs, error handling, or output format. It mentions Grok AI but doesn't explain what that entails operationally, leaving significant gaps in transparency.
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, front-loading the core purpose. It's appropriately sized for the tool's complexity, making it easy to parse without unnecessary elaboration.
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 (4 parameters, no annotations, no output schema), the description is incomplete. It doesn't explain the return values, error conditions, or how function calling integrates with user input, leaving the agent under-informed for effective use.
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%, so the schema fully documents all 4 parameters. The description adds no meaning beyond what the schema provides, not explaining how parameters like messages or tools relate to function calling. Baseline 3 is appropriate as 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 states the tool 'call[s] functions based on user input' using Grok AI, which gives a general purpose but lacks specificity about what functions are called or how this differs from sibling tools like chat_completion. It's vague about the exact verb+resource combination beyond invoking AI capabilities.
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 is provided on when to use this tool versus alternatives like chat_completion or image_understanding. The description implies it's for function calling but doesn't specify contexts, prerequisites, or exclusions, leaving the agent without clear usage direction.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Analyze images using Grok AI vision capabilities (Note: Grok 3 may support image creation)
| Name | Required | Description | Default |
|---|---|---|---|
| base64_image | No | Base64-encoded image data (without the data:image prefix) | |
| image_url | No | URL of the image to analyze | |
| model | No | Grok vision model to use (e.g., grok-2-vision-latest, potentially grok-3 variants) | grok-2-vision-latest |
| prompt | Yes | Text prompt to accompany the image |
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 analyzes images but does not describe what the analysis entails (e.g., object detection, captioning, OCR), potential limitations (e.g., image size restrictions, rate limits), or authentication needs. The note about Grok 3 adds confusion rather than transparency.
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 brief but includes a parenthetical note that is speculative and not directly relevant to the tool's current functionality, reducing efficiency. It is front-loaded with the core purpose, but the extra sentence detracts from conciseness without adding 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?
For a tool with no annotations and no output schema, the description is incomplete. It lacks details on what the analysis returns (e.g., text descriptions, structured data), error conditions, or behavioral traits like rate limits. The note about Grok 3 does not compensate for these gaps, leaving the agent with insufficient context for effective use.
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%, so the schema already documents all four parameters thoroughly. The description adds no additional meaning about parameters beyond what the schema provides, such as explaining interactions between base64_image and image_url or elaborating on model options. Baseline 3 is appropriate as 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 clearly states the tool's purpose as 'Analyze images using Grok AI vision capabilities' with a specific verb ('Analyze') and resource ('images'), distinguishing it from sibling tools like chat_completion and function_calling. However, it includes a parenthetical note about Grok 3 potentially supporting image creation, which slightly dilutes the clarity by introducing unrelated future capabilities.
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 chat_completion or function_calling. It mentions Grok 3 may support image creation, but this is speculative and not actionable for current usage decisions. No explicit when/when-not scenarios or prerequisites are included.
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.
v1.0.0chat_completionfunction_callingimage_understandingScored across 3 tools
Each tool has a clearly distinct purpose: chat_completion handles text generation, function_calling manages function execution, and image_understanding focuses on visual analysis. There is no overlap in functionality, making tool selection straightforward for an agent.
The tools follow a consistent snake_case naming convention, but the pattern is not strictly verb_noun (e.g., chat_completion, function_calling, image_understanding). The naming is readable and logical, with only minor deviations from a perfect pattern.
With only 3 tools, the set feels thin for a general-purpose AI plugin, potentially lacking operations like text summarization, translation, or audio processing. However, it covers core AI functionalities adequately for basic use cases.
The tools cover key AI areas (text, functions, images), but there are notable gaps such as missing text analysis tools (e.g., sentiment analysis, summarization) and no explicit support for audio or video processing. The surface is functional but not comprehensive for a full AI suite.
Share one project context across ChatGPT, Claude, Telegram and any MCP client.
Talk to your public-facing AI from any MCP client — Claude, ChatGPT, Cursor, Cline, Windsurf.
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
Persistent context for Claude. Your AI always knows your projects and next actions across sessions.
