AI Agent Webhook for Document Q&A

The workflow JSON

Copy or download the full n8n JSON below. Paste it into a new n8n workflow, add your credentials, activate. Full import guide →

{
  "nodes": [
    {
      "parameters": {
        "jsonMode": "expressionData",
        "jsonData": "={{ $json.text}}",
        "options": {
          "metadata": {
            "metadataValues": [
              {
                "name": "ticker",
                "value": "={{ $json.ticker }}"
              },
              {
                "name": "quarter",
                "value": "={{ $json.quarter }}"
              },
              {
                "name": "date",
                "value": "={{ $json.date }}"
              },
              {
                "name": "section",
                "value": "={{ $json.section }}"
              },
              {
                "name": "speaker",
                "value": "={{ $json.speaker }}"
              },
              {
                "name": "year",
                "value": "={{ $json.year }}"
              }
            ]
          }
        }
      },
      "type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader",
      "typeVersion": 1.1,
      "position": [
        -656,
        1888
      ],
      "id": "9932b783-659e-43e7-b982-8b4191883c35",
      "name": "Default Data Loader"
    },
    {
      "parameters": {
        "jsCode": "const chunks = $input.first().json.body;\nreturn chunks.map(chunk => ({ json: chunk }));"
      },
      "type": "n8n-nodes-base.code",
      "typeVersion": 2,
      "position": [
        -1072,
        1664
      ],
      "id": "2f5193ad-4fce-42d2-8dc3-8e9188693cc9",
      "name": "Code in JavaScript"
    },
    {
      "parameters": {
        "content": "Used to insert text files to the Pinecone DB (transcripts). Each file has its own metadata of the quarter, year, ticker, speaker. This helps with filtering later on"
      },
      "type": "n8n-nodes-base.stickyNote",
      "typeVersion": 1,
      "position": [
        -1568,
        1648
      ],
      "id": "c4f49a8a-deda-47cd-a2b7-5d451126d8e4",
      "name": "Sticky Note"
    },
    {
      "parameters": {
        "content": "You can send a message of a ticker symbol, ie AMZN",
        "height": 80
      },
      "type": "n8n-nodes-base.stickyNote",
      "typeVersion": 1,
      "position": [
        -1776,
        1296
      ],
      "id": "fe121e63-12e3-49be-9df5-b6589782fb4b",
      "name": "Sticky Note1"
    },
    {
      "parameters": {
        "content": "Call agent using POST request on front-end of website",
        "height": 80
      },
      "type": "n8n-nodes-base.stickyNote",
      "typeVersion": 1,
      "position": [
        -1792,
        1104
      ],
      "id": "17de82bb-9715-4484-a584-5bdbf3746d4c",
      "name": "Sticky Note2"
    },
    {
      "parameters": {
        "mode": "insert",
        "pineconeIndex": {
          "__rl": true,
          "value": "dbmain",
          "mode": "list",
          "cachedResultName": "dbmain"
        },
        "options": {
          "clearNamespace": false,
          "pineconeNamespace": ""
        }
      },
      "id": "1bd5761e-d49b-436e-b3ad-fc750ce35058",
      "name": "Insert into Transcript DB",
      "type": "@n8n/n8n-nodes-langchain.vectorStorePinecone",
      "typeVersion": 1.3,
      "position": [
        -848,
        1664
      ],
      "credentials": {
        "pineconeApi": {
          "name": "<your credential>"
        }
      }
    },
    {
      "parameters": {
        "model": "grok-4-1-fast-non-reasoning",
        "options": {}
      },
      "type": "@n8n/n8n-nodes-langchain.lmChatXAiGrok",
      "typeVersion": 1,
      "position": [
        -1072,
        1344
      ],
      "id": "ade564fa-3498-414d-8742-3f5ad8c59f4e",
      "name": "xAI Grok Chat Model",
      "credentials": {
        "xAiApi": {
          "name": "<your credential>"
        }
      }
    },
    {
      "parameters": {
        "options": {}
      },
      "type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi",
      "typeVersion": 1.2,
      "position": [
        -848,
        1888
      ],
      "id": "436d4b14-f1b3-48a8-9727-e1ebb121d151",
      "name": "Embeddings OpenAI",
      "credentials": {
        "openAiApi": {
          "name": "<your credential>"
        }
      }
    },
    {
      "parameters": {
        "httpMethod": "POST",
        "path": "7d2b11a4-1593-454a-867c-dd7788e07799",
        "options": {}
      },
      "type": "n8n-nodes-base.webhook",
      "typeVersion": 2.1,
      "position": [
        -1296,
        1664
      ],
      "id": "8542537e-1228-4b2d-a01d-b983bb0de8b2",
      "name": "Webhook2"
    },
    {
      "parameters": {
        "options": {}
      },
      "type": "@n8n/n8n-nodes-langchain.chatTrigger",
      "typeVersion": 1.4,
      "position": [
        -1296,
        1328
      ],
      "id": "a16b8058-86cb-4385-8dda-8eefb49dc79c",
      "name": "When chat message received"
    },
    {
      "parameters": {
        "workflowId": {
          "__rl": true,
          "value": "zYQtlFnrHsJQu3Wv",
          "mode": "list",
          "cachedResultUrl": "/workflow/zYQtlFnrHsJQu3Wv",
          "cachedResultName": "Parallel sub-workflow"
        },
        "workflowInputs": {
          "mappingMode": "defineBelow",
          "value": {
            "requestId_transcript": "={{ $json.requestId_transcript }}",
            "output_ticker": "={{ $json.output.ticker }}",
            "requestId_peers": "={{ $('Organize Input').item.json.requestId_peers }}",
            "output_peers": "={{ $('Organize Input').item.json.output.peers }}"
          },
          "matchingColumns": [
            "requestId_transcript",
            "output_ticker",
            "requestId_peers",
            "output_peers"
          ],
          "schema": [
            {
              "id": "requestId_transcript",
              "displayName": "requestId_transcript",
              "required": false,
              "defaultMatch": false,
              "display": true,
              "canBeUsedToMatch": true,
              "removed": false
            },
            {
              "id": "output_ticker",
              "displayName": "output_ticker",
              "required": false,
              "defaultMatch": false,
              "display": true,
              "canBeUsedToMatch": true,
              "removed": false
            },
            {
              "id": "requestId_peers",
              "displayName": "requestId_peers",
              "required": false,
              "defaultMatch": false,
              "display": true,
              "canBeUsedToMatch": true,
              "removed": false
            },
            {
              "id": "output_peers",
              "displayName": "output_peers",
              "required": false,
              "defaultMatch": false,
              "display": true,
              "canBeUsedToMatch": true,
              "removed": false
            }
          ],
          "attemptToConvertTypes": false,
          "convertFieldsToString": true
        },
        "options": {}
      },
      "type": "n8n-nodes-base.executeWorkflow",
      "typeVersion": 1.2,
      "position": [
        -384,
        1232
      ],
      "name": "Call Parallel sub-workflow",
      "id": "f1203988-7b88-4ebf-a14a-7ce2315d50ff"
    },
    {
      "parameters": {
        "options": {}
      },
      "type": "n8n-nodes-base.respondToWebhook",
      "typeVersion": 1.5,
      "position": [
        -160,
        1232
      ],
      "id": "6763c2ad-ad1b-455b-9f33-8f934f26e576",
      "name": "Respond to Webhook"
    },
    {
      "parameters": {
        "content": "The technique I was using for parallelization was causing problems with \"Unused Respond to Webhook node found in the workflow\" when using webhook trigger, because it was seeing this not being used directly for the trigger and erroring. To fix, I first have to create a sub-workflow that performs parallel action, then add a respond to webhook at the end to trick it into being happy.",
        "height": 112,
        "width": 608
      },
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -576,
        1072
      ],
      "typeVersion": 1,
      "id": "c6f0306a-f4ef-47af-bd04-670cadc530f1",
      "name": "Sticky Note3"
    },
    {
      "parameters": {
        "url": "={{ /*n8n-auto-generated-fromAI-override*/ $fromAI('URL', ``, 'string') }}",
        "options": {}
      },
      "type": "n8n-nodes-base.httpRequestTool",
      "typeVersion": 4.4,
      "position": [
        -912,
        1344
      ],
      "id": "1cdd659b-4a51-4676-a734-3c8afc7362a9",
      "name": "Identify Available Peers"
    },
    {
      "parameters": {
        "httpMethod": "POST",
        "path": "ed36debd-2f06-448c-b98e-ef138d3f692a",
        "responseMode": "responseNode",
        "options": {}
      },
      "type": "n8n-nodes-base.webhook",
      "typeVersion": 2.1,
      "position": [
        -1296,
        1120
      ],
      "id": "ec76f7e2-93ec-4a28-b782-5914f1354f54",
      "name": "Fontend Webhook"
    },
    {
      "parameters": {
        "promptType": "define",
        "text": "={{ $json.chatInput || $json.body.chatInput }}",
        "hasOutputParser": true,
        "options": {
          "systemMessage": "You are an equity research assistant. Given a stock ticker, identify the 3 closest publicly traded U.S. peers by business model and industry.\nYou will use this URL to perform a GET Request: https://api.github.com/repos/gselez6761/agentic-ai-challenge-data/contents/data/income_statements\nAmong the folders returned by this URL, here are the rules:\nPeer Selection Criteria (in order of priority):\n\nPeers must compete directly with the target company in at least 2 major revenue segments. If a candidate company does not share at least 2 core business lines with the target, it is NOT a valid peer.\nPeers should be in the same industry or sector. A restaurant chain is not a peer to a tech company. A bank is not a peer to a retailer. An oil company is not a peer to a SaaS platform.\nPeers should be comparable in business model \u2014 if the target is a platform/marketplace, pick other platforms/marketplaces. If the target is a subscription SaaS company, pick other subscription SaaS companies.\nPeers should be similar in scale where possible (large-cap to large-cap, mid-cap to mid-cap), but business model overlap is more important than size.\n\nDisqualification Rules:\n\nDo NOT pick a company just because it is large or well-known. McDonald's is not a peer to Amazon.\nDo NOT pick conglomerates or diversified companies unless the target company is one.\nDo NOT pick companies from a completely different industry even if they share one minor business line (e.g., both having \"subscription services\" does not make Netflix a peer to Costco).\nAsk yourself: \"Would an equity research analyst at a bank put these two companies in the same coverage universe?\" If not, do not select it.\n\nReturn ONLY a JSON object, no explanation, no markdown.\nOutput format:\n{\"ticker\": \"<INPUT_TICKER>\", \"peers\": [\"<PEER1>\", \"<PEER2>\", \"<PEER3>\"]}"
        }
      },
      "type": "@n8n/n8n-nodes-langchain.agent",
      "typeVersion": 3.1,
      "position": [
        -1024,
        1120
      ],
      "id": "7bcd7607-fcb4-4b8a-a64b-762b703f290a",
      "name": "Peer Retrieval Agent"
    },
    {
      "parameters": {
        "jsCode": "// for (const item of $input.all()) {\n//   item.json.requestId = $execution.id + '_' + Date.now();\n// }\n\n// return $input.all();\nfor (const item of $input.all()) {\n  item.json.requestId_transcript = $execution.id + '_transcript_' + Date.now();\n  item.json.requestId_peers = $execution.id + '_peers_' + (Date.now() + 1);\n}\nreturn $input.all();"
      },
      "type": "n8n-nodes-base.code",
      "typeVersion": 2,
      "position": [
        -608,
        1232
      ],
      "id": "20be1e0b-7b02-4b61-8750-fde3830484e1",
      "name": "Organize Input"
    },
    {
      "parameters": {
        "jsonSchemaExample": "{\n\t\"ticker\": \"AMZN\",\n\t\"peers\": [\"WMT\", \"SHOP\"]\n}"
      },
      "type": "@n8n/n8n-nodes-langchain.outputParserStructured",
      "typeVersion": 1.3,
      "position": [
        -752,
        1344
      ],
      "id": "103b6f9a-55c0-4e5e-a9a9-928b42a4552c",
      "name": "format_final_json_response"
    }
  ],
  "connections": {
    "Default Data Loader": {
      "ai_document": [
        [
          {
            "node": "Insert into Transcript DB",
            "type": "ai_document",
            "index": 0
          }
        ]
      ]
    },
    "Code in JavaScript": {
      "main": [
        [
          {
            "node": "Insert into Transcript DB",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Insert into Transcript DB": {
      "main": [
        []
      ]
    },
    "xAI Grok Chat Model": {
      "ai_languageModel": [
        [
          {
            "node": "Peer Retrieval Agent",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "Embeddings OpenAI": {
      "ai_embedding": [
        [
          {
            "node": "Insert into Transcript DB",
            "type": "ai_embedding",
            "index": 0
          }
        ]
      ]
    },
    "Webhook2": {
      "main": [
        [
          {
            "node": "Code in JavaScript",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "When chat message received": {
      "main": [
        [
          {
            "node": "Peer Retrieval Agent",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Call Parallel sub-workflow": {
      "main": [
        [
          {
            "node": "Respond to Webhook",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Identify Available Peers": {
      "ai_tool": [
        [
          {
            "node": "Peer Retrieval Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "Fontend Webhook": {
      "main": [
        [
          {
            "node": "Peer Retrieval Agent",
            "type": "main",
            "index": 0
          }
        ],
        []
      ]
    },
    "Peer Retrieval Agent": {
      "main": [
        [
          {
            "node": "Organize Input",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Organize Input": {
      "main": [
        [
          {
            "node": "Call Parallel sub-workflow",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "format_final_json_response": {
      "ai_outputParser": [
        [
          {
            "node": "Peer Retrieval Agent",
            "type": "ai_outputParser",
            "index": 0
          }
        ]
      ]
    }
  }
}

Credentials you'll need

Each integration node will prompt for credentials when you import. We strip credential IDs before publishing — you'll add your own.

• googlePalmApi
• qdrantApi
• qdrantRestApi
• upstageApi

How this works

This workflow processes incoming documents through a webhook, uses an AI agent to understand queries and retrieve relevant context from a Qdrant vector store, then returns accurate answers based on the loaded material. It suits teams that need reliable document search and question-answering without building their own retrieval pipeline. The key step is the agent’s interaction with Upstage embeddings and the Qdrant vector store, which keeps answers grounded in the supplied data.

Use it when you receive documents via webhook and want fast, context-aware responses; avoid it for simple keyword searches or when you need full manual control over every retrieval step. Common variations include swapping the vector store or adjusting the text splitter for different document lengths.

About this workflow

main_workflow. Uses agent, n8n-nodes-upstage, documentDefaultDataLoader, textSplitterCharacterTextSplitter. Webhook trigger; 63 nodes.

Source: https://github.com/pioneer0413/TZC/blob/c8304cc1bb6e52fa07d9077725c8744b1d541c42/n8n/main_workflow.json — original creator credit. Request a take-down →

More AI & RAG workflows → · Browse all categories →

Related workflows

Workflows that share integrations, category, or trigger type with this one. All free to copy and import.