Context7 Smithery AI Agent MCP Server

• •Clé API Google Gemini

{
  "id": "vqujIu65SFbZUz0n",
  "meta": {
    "instanceId": "3612a0c89f35a354ddd1216385240db51b4635fde7cae2a0d5004be915b832f3",
    "templateCredsSetupCompleted": true
  },
  "name": "Context7 Smithery AI Agent MCP Server",
  "tags": [],
  "nodes": [
    {
      "id": "a42ae423-a348-460e-888c-f3df2b82021d",
      "name": "Modèle de chat Google Gemini",
      "type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
      "position": [
        440,
        380
      ],
      "parameters": {
        "options": {},
        "modelName": "models/gemini-2.5-flash-preview-05-20"
      },
      "credentials": {
        "googlePalmApi": {
          "id": "c8jj7D6oz6RJ81qp",
          "name": "Google Gemini Context7"
        }
      },
      "typeVersion": 1
    },
    {
      "id": "22f3e783-4310-4d15-a875-cbb400a82c1b",
      "name": "context7-resolve-library-id",
      "type": "n8n-nodes-mcp.mcpClientTool",
      "position": [
        1280,
        480
      ],
      "parameters": {
        "toolName": "resolve-library-id",
        "operation": "executeTool",
        "connectionType": "http",
        "toolParameters": "={{ /*n8n-auto-generated-fromAI-override*/ $fromAI('Tool_Parameters', `Generate a JSON object with the parameters for this tool.\nThe tool expects a JSON object with one **required** key:\n- \"libraryName\": (string) The common name of the software library or framework.\nExample: {\"libraryName\": \"react\"}`, 'json') }}",
        "descriptionType": "manual",
        "toolDescription": "Resolves a general library name (e.g., 'nextjs', 'mongodb') into a unique Context7-compatible library ID (e.g., '/vercel/next.js', '/mongodb/docs'). Use this tool first to get the correct ID before attempting to fetch its documentation."
      },
      "credentials": {
        "mcpClientHttpApi": {
          "id": "CL3sgNzCgbtRRxtL",
          "name": "smithery_context7"
        }
      },
      "typeVersion": 1
    },
    {
      "id": "526eab1c-7a08-45eb-8169-7e8d4578c117",
      "name": "context7-get-library-docs",
      "type": "n8n-nodes-mcp.mcpClientTool",
      "position": [
        1840,
        480
      ],
      "parameters": {
        "toolName": "get-library-docs",
        "operation": "executeTool",
        "connectionType": "http",
        "toolParameters": "={{ /*n8n-auto-generated-fromAI-override*/ $fromAI('Tool_Parameters', `Generate a JSON object with the parameters for this tool.\nThe tool expects a JSON object with the following keys:\n- \"context7CompatibleLibraryID\": (string, **required**) The exact Context7-compatible library ID (e.g., \"/vercel/next.js\").\n- \"topic\": (string, optional) A specific topic to focus the documentation on (e.g., \"routing\").\n- \"tokens\": (number, optional, default: 10000) Maximum number of tokens for the documentation. Minimum effective value is 10000.\nExample for required only: {\"context7CompatibleLibraryID\": \"/mongodb/docs\"}\nExample with optional: {\"context7CompatibleLibraryID\": \"/vercel/next.js\", \"topic\": \"authentication\", \"tokens\": 15000}`, 'json') }}",
        "descriptionType": "manual",
        "toolDescription": "Fetches documentation for a specified library using its Context7-compatible library ID. You can optionally focus on a specific topic within the documentation and set a maximum token limit for the returned content."
      },
      "credentials": {
        "mcpClientHttpApi": {
          "id": "CL3sgNzCgbtRRxtL",
          "name": "smithery_context7"
        }
      },
      "typeVersion": 1
    },
    {
      "id": "5cafb0d4-19ac-4589-88d8-34b88404e0f1",
      "name": "Context7 MCP Server Trigger",
      "type": "@n8n/n8n-nodes-langchain.mcpTrigger",
      "position": [
        120,
        -860
      ],
      "webhookId": "c597b9e9-33bf-41d9-8602-e63d2bee174b",
      "parameters": {
        "path": "dafdfsdfsdf6e8-ea19-43b7-b337-a5iuodsf98vjfewf"
      },
      "typeVersion": 1
    },
    {
      "id": "945d8818-c346-44ba-b4c1-56d166a4f381",
      "name": "Context7 Agent IA",
      "type": "@n8n/n8n-nodes-langchain.agent",
      "position": [
        880,
        -20
      ],
      "parameters": {
        "text": "={{ $('Context7 Workflow Start').item.json.query }}",
        "options": {
          "systemMessage": "You are an AI assistant with access to Context7 tools to help users find software library documentation.\nYour primary functions are to:\n1.  **Resolve Library IDs:** When a user asks for documentation for a library like 'Next.js' or 'MongoDB', first use the `resolve-library-id` tool to find its specific Context7-compatible ID (e.g., '/vercel/next.js'). The input for this tool is the `libraryName`.\n2.  **Get Library Documentation:** Once you have the `context7CompatibleLibraryID` from the previous step, use the `get-library-docs` tool to fetch the documentation.\n    *   The required parameter is `context7CompatibleLibraryID`.\n    *   If the user specifies a particular `topic` (e.g., \"routing in Next.js\"), include this in the `topic` parameter.\n    *   You can also adjust the `tokens` parameter if a different length of documentation is implicitly or explicitly requested, keeping in mind the default and minimum is 10000.\n\nAlways use `resolve-library-id` before `get-library-docs` if you don't already have the exact Context7-compatible library ID. Be precise with the IDs."
        },
        "promptType": "define"
      },
      "typeVersion": 2
    },
    {
      "id": "1b0cd3cc-8566-4cb5-932c-005149a2467b",
      "name": "Context7 Workflow Début",
      "type": "n8n-nodes-base.executeWorkflowTrigger",
      "position": [
        100,
        -20
      ],
      "parameters": {
        "workflowInputs": {
          "values": [
            {
              "name": "query"
            }
          ]
        }
      },
      "typeVersion": 1.1
    },
    {
      "id": "e955aa60-2b80-45d5-885b-24b7fe56a426",
      "name": "Simple Mémoire",
      "type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
      "position": [
        700,
        520
      ],
      "parameters": {
        "sessionKey": "={{ $execution.id }}",
        "sessionIdType": "customKey"
      },
      "typeVersion": 1.3
    },
    {
      "id": "4927c593-5762-4a26-87eb-8a7e0a4b3d32",
      "name": "call_context7_ai_agent",
      "type": "@n8n/n8n-nodes-langchain.toolWorkflow",
      "position": [
        280,
        -460
      ],
      "parameters": {
        "workflowId": {
          "__rl": true,
          "mode": "list",
          "value": "vqujIu65SFbZUz0n",
          "cachedResultName": "Context7 Smithery AI Agent MCP Server"
        },
        "description": "=Processes a natural language query to find and retrieve software library documentation using Context7. It intelligently uses sub-tools to first resolve library names to Context7 IDs and then fetches the relevant documentation.",
        "workflowInputs": {
          "value": {
            "query": "={{ /*n8n-auto-generated-fromAI-override*/ $fromAI('query', `The user's natural language question or request for software library documentation. For example: 'How do I implement authentication in Next.js?' or 'Get docs for MongoDB aggregation framework'.`, 'string') }}"
          },
          "schema": [
            {
              "id": "query",
              "type": "string",
              "display": true,
              "removed": false,
              "required": false,
              "displayName": "query",
              "defaultMatch": false,
              "canBeUsedToMatch": true
            }
          ],
          "mappingMode": "defineBelow",
          "matchingColumns": [],
          "attemptToConvertTypes": false,
          "convertFieldsToString": false
        }
      },
      "typeVersion": 2.2
    },
    {
      "id": "d72a2661-5622-4bfa-9d42-8c96b5fef877",
      "name": "Note adhésive",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1700,
        -40
      ],
      "parameters": {
        "width": 680,
        "height": 320,
        "content": "## MCP JSON for Roo Code etc\n\nTo use this in Roo Code, Cline etc use\n\n\"n8n-context7\": {\n      \"url\": \"https://n8n.coolify.au/mcp/dafd56e8-ea19-43b7-b337-a5881c3bc541/sse\",\n      \"alwaysAllow\": [\n        \"call_context7_ai_agent\"\n      ]\n    }\n\nwhere the url is the the path that you get from the mcp server node in n8n"
      },
      "typeVersion": 1
    },
    {
      "id": "c0dc815b-3d0f-49e2-8da0-08d2f45d0991",
      "name": "Note adhésive1",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -260,
        -1140
      ],
      "parameters": {
        "color": 4,
        "width": 580,
        "height": 420,
        "content": "This node is the public entry point for the Context7 AI Agent service.\nIt exposes an MCP (Multi-Agent Collaboration Protocol) SSE (Server-Sent Events) endpoint.\n\n- Path: The long, random path (e.g., /dafdfsdfsdf... ) acts as a simple 'security by obscurity' measure.\n- Function: It listens for incoming HTTP POST requests to its SSE endpoint.\n- Tool Exposed: This trigger makes the 'call_context7_ai_agent' tool (defined by the connected Tool Workflow node) available to external MCP clients.\n- Input: Expects a JSON payload from the client, typically like: {\"input\": {\"query\": \"user's question\"}}\n- Output: Streams SSE events back to the client, containing the AI agent's responses and intermediate steps."
      },
      "typeVersion": 1
    },
    {
      "id": "adc54a99-b04d-40be-a9a4-25f837453c14",
      "name": "Note adhésive2",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -260,
        -680
      ],
      "parameters": {
        "color": 4,
        "width": 580,
        "height": 340,
        "content": "This node represents the tool that external clients call via the MCP Server Trigger.\nWhen triggered, it executes the separate \"Context7 Smithery AI Agent MCP Server\" sub-workflow (ID: vqujIu65SFbZUz0n).\n\n- Action: Invokes the AI agent sub-workflow to process the user's request.\n- Inputs to Sub-Workflow:\n    - 'query': The user's natural language question, received from the external MCP client. This is passed directly to the sub-workflow.\n- Output: The result from the sub-workflow (the AI agent's final response or stream) is passed back through the MCP Server Trigger to the client."
      },
      "typeVersion": 1
    },
    {
      "id": "5cc037d1-4155-49dc-86f5-7af9976149bf",
      "name": "Note adhésive3",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1700,
        -660
      ],
      "parameters": {
        "color": 2,
        "width": 680,
        "height": 600,
        "content": "## Main MCP Server Workflow Overview\n\nThis workflow acts as the public gateway for the Context7 AI Documentation Agent.\n\n1.  **`Context7 MCP Server Trigger`**:\n    *   Receives requests from external clients (e.g., Roo Code, Cline) at its unique SSE endpoint URL.\n    *   The path is intentionally long and random for basic access control.\n    *   Client JSON for Roo Code etc.:\n        \"n8n-context7\": {\n          \"url\": \"https://n8n.coolify.au/mcp/YOUR_LONG_RANDOM_PATH/sse\",\n          \"alwaysAllow\": [\n            \"call_context7_ai_agent\"  // This is the tool name defined in the Tool Workflow node\n          ]\n        }\n        (Replace YOUR_LONG_RANDOM_PATH with the actual path from the trigger node)\n\n2.  **`call_context7_ai_agent` (Tool Workflow Node)**:\n    *   This is the \"tool\" the MCP Server Trigger exposes.\n    *   It calls the separate AI Agent Sub-Workflow (ID: vqujIu65SFbZUz0n), passing the user's `query`.\n\n**Data Flow:**\nExternal Client -> MCP Server Trigger -> call_context7_ai_agent (Tool) -> AI Agent Sub-Workflow -> (Response Stream) -> MCP Server Trigger -> External Client\n\n3.  ** There is a video I've made on youtube that explains this workflow a bit and why I created it: https://youtu.be/dudvmyp7Pyg?si=N59DYTe2WE6GPlvf"
      },
      "typeVersion": 1
    },
    {
      "id": "79bf63c1-b325-4a3c-950e-30df5aef15e4",
      "name": "Note adhésive4",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -20,
        -200
      ],
      "parameters": {
        "color": 5,
        "width": 360,
        "height": 200,
        "content": "This is the entry point for the AI Agent sub-workflow.\n- Receives:\n    - 'query': The user's natural language question from the main workflow.\n- Passes data to: The IF node (for API key check, if implemented) or directly to the AI Agent."
      },
      "typeVersion": 1
    },
    {
      "id": "2486cacc-b06a-4e30-b363-456930cf0ea5",
      "name": "Note adhésive5",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        760,
        -280
      ],
      "parameters": {
        "color": 5,
        "width": 560,
        "height": 440,
        "content": "The core AI logic unit.\n- Model: Uses Google Gemini.\n- System Prompt: Instructs the AI on how to use the Context7 tools.\n- Tools: Has access to 'context7-resolve-library-id' and 'context7-get-library-docs'.\n- Memory: Connected to 'Simple Memory' using the current execution ID ({{ $execution.id }}) as the session key for conversation context.\n- Input: Receives the user's 'query'.\n- Output: Generates a response by orchestrating calls to its tools based on the query."
      },
      "typeVersion": 1
    },
    {
      "id": "f5f12aa6-d90a-419c-9baa-2c7c451a71d1",
      "name": "Note adhésive6",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        580,
        640
      ],
      "parameters": {
        "color": 6,
        "width": 360,
        "height": 200,
        "content": "Provides conversational memory for the AI Agent.\n- Type: Window Buffer Memory (keeps a sliding window of recent messages).\n- Session Key: Uses `{{ $execution.id }}`. This ensures that memory is unique to each individual execution of this sub-workflow, preventing cross-talk between different user requests."
      },
      "typeVersion": 1
    },
    {
      "id": "d2598e8e-3931-436f-a480-b4bef659bccb",
      "name": "Note adhésive7",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1100,
        640
      ],
      "parameters": {
        "color": 6,
        "width": 480,
        "height": 500,
        "content": "This node allows the AI Agent to use the external 'resolve-library-id' Context7 MCP tool.\n\n- Connection: Uses the 'smithery_context7' credential to communicate with the Context7 MCP service.\n- Action: When invoked by the AI Agent, this node makes a call to the Context7 'resolve-library-id' tool.\n- Purpose of External Tool: To convert a common library name (e.g., \"react\") into a Context7-specific ID (e.g., \"/facebook/react\").\n- AI Agent Interaction:\n    - The AI Agent's system prompt guides it to use this tool first to get a library ID.\n    - The \"Tool Parameters\" for this node are configured to use an AI-generated JSON object. The AI is prompted with: \"Generate a JSON object with the parameters for this tool. The tool expects a JSON object with one **required** key: - \"libraryName\": (string) The common name of the software library or framework. Example: {\\\"libraryName\\\": \\\"react\\\"}\"\n    - The AI Agent will construct this JSON (e.g., `{\"libraryName\": \"nextjs\"}`) and pass it to this node.\n- Output: Returns the Context7-compatible library ID to the AI Agent."
      },
      "typeVersion": 1
    },
    {
      "id": "7bebd852-75e5-427e-a3ee-781cd85f20e6",
      "name": "Note adhésive8",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1620,
        640
      ],
      "parameters": {
        "color": 6,
        "width": 520,
        "height": 500,
        "content": "This node allows the AI Agent to use the external 'get-library-docs' Context7 MCP tool.\n\n- Connection: Uses the 'smithery_context7' credential to communicate with the Context7 MCP service.\n- Action: When invoked by the AI Agent, this node makes a call to the Context7 'get-library-docs' tool.\n- Purpose of External Tool: To fetch documentation for a library, given its Context7-specific ID.\n- AI Agent Interaction:\n    - The AI Agent's system prompt guides it to use this tool after obtaining a `context7CompatibleLibraryID` (usually from the 'resolve-library-id' tool).\n    - The \"Tool Parameters\" for this node are configured to use an AI-generated JSON object. The AI is prompted with details about expected keys like `context7CompatibleLibraryID` (required), `topic` (optional), and `tokens` (optional).\n    - Example JSON the AI might construct: `{\"context7CompatibleLibraryID\": \"/vercel/next.js\", \"topic\": \"routing\"}`.\n- Output: Returns the fetched library documentation (or relevant snippets) to the AI Agent."
      },
      "typeVersion": 1
    },
    {
      "id": "f4749f9a-6f9e-45c5-b6a9-a89d4dc77dec",
      "name": "Note adhésive9",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        500,
        -1380
      ],
      "parameters": {
        "color": 7,
        "width": 1140,
        "height": 1040,
        "content": "## n8n AI Agent as an MCP Service (Context7 Docs Example)\n\n**What This Workflow Does:**\nThis n8n setup transforms a complex AI agent into a single, callable tool exposed over an MCP (Multi-Agent Collaboration Protocol) SSE (Server-Sent Events) endpoint. It allows external clients (like development environments, other AI agents, or custom applications) to leverage the power of an n8n-hosted AI agent with a simple natural language query.\n\n**How It Works:**\n1.  **`Context7 MCP Server Trigger` (This Workflow):** Creates a public SSE endpoint. Its path is kept long and random for basic security by obscurity. This trigger exposes the `call_context7_ai_agent` tool.\n2.  **`call_context7_ai_agent` (Tool Workflow Node, This Workflow):** This node is the \"tool\" offered by the MCP server. When called by a client, it doesn't do the AI work itself; instead, it invokes a separate sub-workflow.\n3.  **AI Agent Sub-Workflow (e.g., \"Context7 Smithery AI Agent MCP Server\" - ID: vqujIu65SFbZUz0n):** This is where the magic happens!\n    *   It receives the `query` from the main workflow.\n    *   A `Context7 AI Agent` node, powered by a model like Google Gemini, orchestrates the task.\n    *   The agent uses connected tools:\n        *   `context7-resolve-library-id`: To find the correct ID for a software library.\n        *   `context7-get-library-docs`: To fetch documentation for that library.\n    *   It uses `Simple Memory` (scoped by execution ID) to maintain conversational context if needed for multi-turn interactions (though this example is primarily single-shot).\n\n**Why This Pattern is Useful & The Smithery Connection:**\n\n*   **Simplifies Client-Side Logic:** Clients (like **Smithery's Roo code**, VS Code extensions, or other AI systems) don't need to manage complex multi-step reasoning, tool selection, or API calls to different services. They just call *one tool* (`call_context7_ai_agent`) with a plain language request (e.g., \"What is flexbox in Tailwind CSS?\").\n*   **Reduces Client-Side Token Usage:** All the detailed system prompts, tool descriptions for the sub-tools (resolve-id, get-docs), and the actual execution of these tools happen *within n8n*. The client only sends a short query and receives the processed result. This is incredibly efficient, especially if the client is another LLM.\n*   **Centralized Agent Capabilities:** You can update, enhance, or swap out models and tools within your n8n AI agent sub-workflow without needing to update any of the client applications that consume it.\n*   **Ideal for Agentic Systems:** This pattern is a cornerstone of building robust **agentic systems**. By encapsulating specialized AI capabilities within n8n and exposing them as callable MCP tools:\n    *   You can create a \"team\" of specialized n8n agents.\n    *   A primary agent (in Smithery, or another n8n workflow) can delegate tasks to these specialized n8n agent-tools.\n    *   It promotes modularity, reusability, and scalability in AI application development.\n*   **Leverage n8n's Ecosystem:** Build your agent with n8n's visual interface, manage credentials securely, and easily integrate with hundreds of other services if your agent needs them.\n\n**Client Configuration Example (for Roo Code/Cline):**\nTo connect to this n8n agent from a client like Roo code, you'd configure it with the MCP server details:\n```json\n\"n8n-context7-agent\": { // A descriptive name for your client\n  \"url\": \"https://YOUR_N8N_INSTANCE/mcp/YOUR_LONG_RANDOM_PATH/sse\", // Path from MCP Server Trigger\n  \"alwaysAllow\": [\n    \"call_context7_ai_agent\" // Name of the Tool Workflow node\n  ]\n}"
      },
      "typeVersion": 1
    }
  ],
  "active": true,
  "pinData": {},
  "settings": {
    "executionOrder": "v1"
  },
  "versionId": "785e4ebf-bb80-48e3-872c-40d1b9a8d14b",
  "connections": {
    "Simple Memory": {
      "ai_memory": [
        [
          {
            "node": "Context7 AI Agent",
            "type": "ai_memory",
            "index": 0
          }
        ]
      ]
    },
    "4927c593-5762-4a26-87eb-8a7e0a4b3d32": {
      "ai_tool": [
        [
          {
            "node": "5cafb0d4-19ac-4589-88d8-34b88404e0f1",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "Context7 Workflow Start": {
      "main": [
        [
          {
            "node": "Context7 AI Agent",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Google Gemini Chat Model": {
      "ai_languageModel": [
        [
          {
            "node": "Context7 AI Agent",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "526eab1c-7a08-45eb-8169-7e8d4578c117": {
      "ai_tool": [
        [
          {
            "node": "Context7 AI Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "22f3e783-4310-4d15-a875-cbb400a82c1b": {
      "ai_tool": [
        [
          {
            "node": "Context7 AI Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    }
  }
}