|<[^>]+>[^<]*<\\/[^>]+>/g) }}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "6ce79f16-51e3-4192-8103-738222be558b",
"name": "Notion Block Generator",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
100,
1600
],
"parameters": {
"text": "={{ $json.tag.trim() }}",
"messages": {
"messageValues": [
{
"message": "=Convert the following html into its equivalent Notion Block as per Notion's API schema.\n* Ensure the content is always included and remains the same.\n* Return only a json response.\n* Generate child-level blocks. Should not define \"parent\" or \"children\" property.\n* Strongly prefer headings, paragraphs, tables and lists type blocks.\n* available headings are heading_1, heading_2 and heading_3 - h4,h5,h6 should use heading_3 type instead. ensure headings use the rich text definition.\n* ensure lists blocks include all list items.\n\n## Examples\n\n1. headings\n```\nReferences
\n```\nwould convert to \n```\n{\"object\": \"block\", \"type\": \"heading_3\", \"heading_3\": { \"rich_text\": [{\"type\": \"text\",\"text\": {\"content\": \"References\"}}]}}\n```\n\n2. lists\n```\n\n```\nwould convert to\n```\n[\n{\n \"object\": \"block\",\n \"type\": \"bulleted_list_item\",\n \"bulleted_list_item\": {\"rich_text\": [{\"type\": \"text\",\"text\": {\"content\": \"hello\"}}]}\n},\n{\n \"object\": \"block\",\n \"type\": \"bulleted_list_item\",\n \"bulleted_list_item\": {\"rich_text\": [{\"type\": \"text\",\"text\": {\"content\": \"world\"}}]}\n}\n]\n```\n\n3. tables\n```\n\n \n | Technology | Potential Impact |
|---|
\n \n \n \n | 5G Connectivity | Enables faster data speeds and advanced apps | \n
\n \n
\n```\nwould convert to\n```\n{\n \"object\": \"block\",\n \"type\": \"table\",\n \"table\": {\n \"table_width\": 2,\n \"has_column_header\": true,\n \"has_row_header\": false,\n \"children\": [\n {\n \"object\": \"block\",\n \"type\": \"table_row\",\n \"table_row\": {\n \"cells\": [\n [\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \"Technology\",\n \"link\": null\n }\n },\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \"Potential Impact\",\n \"link\": null\n }\n }\n ],\n [\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \"5G Connectivity\",\n \"link\": null\n }\n },\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \"Enables faster data speeds and advanced apps\",\n \"link\": null\n }\n }\n ]\n ]\n }\n }\n ]\n }\n}\n```\n4. anchor links\nSince Notion doesn't support anchor links, just convert them to rich text blocks instead.\n```\nModule 0: Pre-Course Setup and Learning Principles\n```\nconverts to\n```\n{\n \"object\": \"block\",\n \"type\": \"paragraph\",\n \"paragraph\": {\n \"rich_text\": [\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \"Module 0: Pre-Course Setup and Learning Principles\"\n }\n }\n ]\n }\n}\n```\n5. Invalid html parts\nWhen the html is not syntax valid eg. orphaned closing tags, then just skip the conversion and use an empty rich text block.\n```\n\\n\n```\ncan be substituted with\n```\n{\n \"object\": \"block\",\n \"type\": \"paragraph\",\n \"paragraph\": {\n \"rich_text\": [\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \" \"\n }\n }\n ]\n }\n}\n```"
}
]
},
"promptType": "define"
},
"typeVersion": 1.5
},
{
"id": "e3eeb9f0-7407-41f9-a814-def6c26b2ee1",
"name": "Google Gemini Chat Model",
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"position": [
80,
1760
],
"parameters": {
"options": {},
"modelName": "models/gemini-2.0-flash"
},
"credentials": {
"googlePalmApi": {
"id": "dSxo6ns5wn658r8N",
"name": "Google Gemini(PaLM) Api account"
}
},
"typeVersion": 1
},
{
"id": "5b0aeaca-dce5-4afd-8a8a-0ef2c18b6f06",
"name": "Parse JSON blocks",
"type": "n8n-nodes-base.set",
"onError": "continueRegularOutput",
"position": [
420,
1600
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "73fcb8a0-2672-4bd5-86de-8075e1e02baf",
"name": "=block",
"type": "array",
"value": "={{\n(function(){\n const block = $json.text\n .replace('```json', '')\n .replace('```', '')\n .trim()\n .parseJson();\n if (Array.isArray(block)) return block;\n if (block.type.startsWith('heading_')) {\n const prev = Number(block.type.split('_')[1]);\n const next = Math.max(1, prev - 1);\n if (next !== prev) {\n block.type = `heading_${next}`;\n block[`heading_${next}`] = Object.assign({}, block[`heading_${prev}`]);\n block[`heading_${prev}`] = undefined;\n }\n }\n return [block];\n})()\n}}"
}
]
}
},
"executeOnce": false,
"typeVersion": 3.4
},
{
"id": "e2a5a5bc-a3c8-42c5-9419-74ce3525f599",
"name": "Upload to Notion Page",
"type": "n8n-nodes-base.httpRequest",
"onError": "continueRegularOutput",
"maxTries": 2,
"position": [
1680,
1760
],
"parameters": {
"url": "=https://api.notion.com/v1/blocks/{{ $('Get Existing Row1').first().json.id }}/children",
"method": "PATCH",
"options": {
"timeout": "={{ 1000 * 60 }}"
},
"jsonBody": "={{\n{\n \"children\": $json.block\n}\n}}",
"sendBody": true,
"sendHeaders": true,
"specifyBody": "json",
"authentication": "predefinedCredentialType",
"headerParameters": {
"parameters": [
{
"name": "Notion-Version",
"value": "2022-06-28"
}
]
},
"nodeCredentialType": "notionApi"
},
"credentials": {
"notionApi": {
"id": "iHBHe7ypzz4mZExM",
"name": "Notion account"
}
},
"retryOnFail": true,
"typeVersion": 4.2,
"waitBetweenTries": 3000
},
{
"id": "9f18b2a5-ba74-40fc-8e35-a93ecd13507a",
"name": "Sticky Note5",
"type": "n8n-nodes-base.stickyNote",
"position": [
-520,
620
],
"parameters": {
"color": 7,
"width": 1740,
"height": 740,
"content": "## 8. Web Search and Extracting Web Page Contents using [APIFY.com](https://www.apify.com?fpr=414q6)\n[Read more about the HTTP Request node](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.httprequest/)\n\nHere is where I deviated a little from the reference implementation. I opted not to use Firecrawl.ai due to (1) high cost of the service and (2) a regular non-ai crawler would work just as well and probably quicker.\nInstead I'm using [APIFY.com](https://www.apify.com?fpr=414q6) which is a more performant, cost-effective and reliable web scraper service. If you don't want to use Apify, feel free to swap this out with your preferred service.\n\nThis step is the most exciting in terms of improvements and optimisations eg. mix in internal data sources! Add in Perplexity.ai or Jina.ai! Possibilities are endless."
},
"typeVersion": 1
},
{
"id": "84c34a2a-d8bb-4e62-a5ea-df0a142aa2b4",
"name": "Sticky Note6",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1140,
60
],
"parameters": {
"color": 7,
"width": 460,
"height": 360,
"content": "## 5. Set Report to In-Progress\n[Read more about the Notion node](https://docs.n8n.io/integrations/builtin/app-nodes/n8n-nodes-base.notion/)"
},
"typeVersion": 1
},
{
"id": "bd022636-873d-4aca-8929-c189f8596cc1",
"name": "Sticky Note7",
"type": "n8n-nodes-base.stickyNote",
"position": [
1240,
700
],
"parameters": {
"color": 7,
"width": 780,
"height": 660,
"content": "## 9. Compile Learnings with Reasoning Model\n[Read more about the Basic LLM node](https://docs.n8n.io/integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainllm/)\n\nWith our gathered sources, it's now just a case of giving it to our LLM to compile a list of \"learnings\" from them. For our DeepResearcher, we'll use OpenAI's o3-mini which is the latest reasoning model at time of writing. Reasoning perform better than regular chat models due their chain-of-thought or \"thinking\" process that they perform.\n\nThe \"Learnings\" are then combined with the generated research goal to complete one loop."
},
"typeVersion": 1
},
{
"id": "b2f6e51d-cbe6-4459-9515-679f79063926",
"name": "Get Existing Row1",
"type": "n8n-nodes-base.notion",
"position": [
-1020,
1600
],
"parameters": {
"limit": 1,
"filters": {
"conditions": [
{
"key": "Request ID|rich_text",
"condition": "equals",
"richTextValue": "={{ $json.requestId.toString() }}"
}
]
},
"options": {},
"resource": "databasePage",
"matchType": "allFilters",
"operation": "getAll",
"databaseId": {
"__rl": true,
"mode": "list",
"value": "19486dd6-0c0c-80da-9cb7-eb1468ea9afd",
"cachedResultUrl": "https://www.notion.so/19486dd60c0c80da9cb7eb1468ea9afd",
"cachedResultName": "n8n DeepResearch"
},
"filterType": "manual"
},
"credentials": {
"notionApi": {
"id": "iHBHe7ypzz4mZExM",
"name": "Notion account"
}
},
"typeVersion": 2.2
},
{
"id": "7b315060-7e40-410c-ac9d-ef22acbb175a",
"name": "Sticky Note8",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1140,
1400
],
"parameters": {
"color": 7,
"width": 660,
"height": 540,
"content": "## 10. Generate DeepSearch Report using Learnings\n[Read more about the Basic LLM node](https://docs.n8n.io/integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainllm/)\n\nFinally! After all learnings have been gathered - which may have taken up to an hour or more on the higher settings! - they are given to our LLM to generate the final research report in markdown format. Technically, the DeepResearch ends here but for this template, we need to push the output to Notion. If you're not using Notion, feel free to ignore the last few steps."
},
"typeVersion": 1
},
{
"id": "01aaf2a8-c145-4f39-aa52-f40fc28f8767",
"name": "Sticky Note9",
"type": "n8n-nodes-base.stickyNote",
"position": [
-460,
1400
],
"parameters": {
"color": 7,
"width": 1060,
"height": 540,
"content": "## 11. Reformat Report as Notion Blocks\n[Learn more about the Markdown node](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.markdown/)\n\nTo write our report to our Notion page, we'll have to convert it to Notion \"blocks\" - these are specialised json objects which are required by the Notion API. There are quite a number of ways to do this conversion not involving the use of AI but for kicks, I decided to do so anyway. In this step, we first convert to HTML as it allows us to split the report semantically and makes for easier parsing for the LLM."
},
"typeVersion": 1
},
{
"id": "700a6f44-86bf-4aab-8a42-23bf6843f681",
"name": "Sticky Note10",
"type": "n8n-nodes-base.stickyNote",
"position": [
1220,
1400
],
"parameters": {
"color": 7,
"width": 800,
"height": 580,
"content": "## 13. Update Report in Notion\n[Read more about the HTTP request node](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.httprequest/)\n\nIn this step, we can use the Notion API to add the blocks to our page sequentially. A loop is used due to the unstable Notion API - the loop allows retries for blocks that require it."
},
"typeVersion": 1
},
{
"id": "f8536052-c851-42ec-aaf3-5fc876570f6d",
"name": "Sticky Note11",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1840,
-680
],
"parameters": {
"color": 7,
"width": 680,
"height": 560,
"content": "## 1. Let's Research!\n[Learn more about the form trigger node](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.formtrigger)\n\nn8n forms are a really nice way to get our frontend up and running quickly and compared to chat, offers a superior user interface for user input. I've gone perhaps a little extra with the custom html fields but I do enjoy adding a little customisation now and then."
},
"typeVersion": 1
},
{
"id": "58aa92a2-e1fd-497d-a27e-40b733189bab",
"name": "DeepResearch Report",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
-860,
1600
],
"parameters": {
"text": "=You are are an expert and insightful researcher.\n* Given the following prompt from the user, write a final report on the topic using the learnings from research.\n* Make it as as detailed as possible, aim for 3 or more pages, include ALL the learnings from research.\n* Format the report in markdown. Use headings, lists and tables only and where appropriate.\n\n{{ $('JobType Router').first().json.data.query }}\n\nHere are all the learnings from previous research:\n\n\n{{\n$('JobType Router').first().json.data\n .all_learnings\n .map(item => `${item}`) \n .join('\\n')\n}}\n",
"promptType": "define"
},
"typeVersion": 1.5
},
{
"id": "0656be83-d510-46f1-aeeb-f62a69aa3cf2",
"name": "DeepResearch Learnings",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
1500,
980
],
"parameters": {
"text": "=Given the following contents from a SERP search for the query {{ $('Item Ref').first().json.query }}, generate a list of learnings from the contents. Return a maximum of 3 learnings, but feel free to return less if the contents are clear. Make sure each learning is unique and not similar to each other. The learnings should be concise and to the point, as detailed and infromation dense as possible. Make sure to include any entities like people, places, companies, products, things, etc in the learnings, as well as any exact metrics, numbers, or dates. The learnings will be used to research the topic further.\n\n\n{{\n$('Convert to Markdown')\n .all()\n .map(item =>`\\n${item.json.markdown.substr(0, 25_000)}\\n`)\n .join('\\n')\n}}\n",
"messages": {
"messageValues": [
{
"type": "HumanMessagePromptTemplate",
"message": "=You are an expert researcher. Today is {{ $now.toLocaleString() }}. Follow these instructions when responding:\n - You may be asked to research subjects that is after your knowledge cutoff, assume the user is right when presented with news.\n - The user is a highly experienced analyst, no need to simplify it, be as detailed as possible and make sure your response is correct.\n - Be highly organized.\n - Suggest solutions that I didn't think about.\n - Be proactive and anticipate my needs.\n - Treat me as an expert in all subject matter.\n - Mistakes erode my trust, so be accurate and thorough.\n - Provide detailed explanations, I'm comfortable with lots of detail.\n - Value good arguments over authorities, the source is irrelevant.\n - Consider new technologies and contrarian ideas, not just the conventional wisdom.\n - You may use high levels of speculation or prediction, just flag it for me."
}
]
},
"promptType": "define",
"hasOutputParser": true
},
"executeOnce": true,
"typeVersion": 1.5
},
{
"id": "9296a787-3226-44fe-8118-f84dda8e5167",
"name": "Generate Report",
"type": "n8n-nodes-base.executeWorkflow",
"position": [
480,
180
],
"parameters": {
"options": {
"waitForSubWorkflow": false
},
"workflowId": {
"__rl": true,
"mode": "id",
"value": "={{ $workflow.id }}"
},
"workflowInputs": {
"value": {
"data": "={{\n{\n ...Object.assign({}, $json),\n query: $('JobType Router').first().json.data.query\n}\n}}",
"jobType": "deepresearch_report",
"requestId": "={{ $('JobType Router').first().json.requestId }}"
},
"schema": [
{
"id": "requestId",
"display": true,
"required": false,
"displayName": "requestId",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "jobType",
"type": "string",
"display": true,
"required": false,
"displayName": "jobType",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "data",
"type": "object",
"display": true,
"required": false,
"displayName": "data",
"defaultMatch": false,
"canBeUsedToMatch": true
}
],
"mappingMode": "defineBelow",
"matchingColumns": [],
"attemptToConvertTypes": false,
"convertFieldsToString": true
}
},
"typeVersion": 1.2
},
{
"id": "471f9b9f-f331-4652-95de-1ec7136ea692",
"name": "Generate Learnings",
"type": "n8n-nodes-base.executeWorkflow",
"position": [
-380,
180
],
"parameters": {
"mode": "each",
"options": {
"waitForSubWorkflow": true
},
"workflowId": {
"__rl": true,
"mode": "id",
"value": "={{ $workflow.id }}"
},
"workflowInputs": {
"value": {
"data": "={{ $json }}",
"jobType": "deepresearch_learnings",
"requestId": "={{ $('JobType Router').first().json.requestId }}"
},
"schema": [
{
"id": "requestId",
"display": true,
"removed": false,
"required": false,
"displayName": "requestId",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "jobType",
"type": "string",
"display": true,
"removed": false,
"required": false,
"displayName": "jobType",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "data",
"type": "object",
"display": true,
"removed": false,
"required": false,
"displayName": "data",
"defaultMatch": false,
"canBeUsedToMatch": true
}
],
"mappingMode": "defineBelow",
"matchingColumns": [],
"attemptToConvertTypes": false,
"convertFieldsToString": true
}
},
"typeVersion": 1.2
},
{
"id": "2e2fa1e6-9d5c-46ff-985c-58ada1139837",
"name": "Confirmation",
"type": "n8n-nodes-base.form",
"position": [
780,
-420
],
"webhookId": "cf41a176-5d30-4274-955e-b0d5b483d37f",
"parameters": {
"options": {
"formTitle": "DeepResearcher",
"buttonLabel": "Done",
"formDescription": "=\n\nYour Report Is On Its Way!\n
\nDeepResearcher will now work independently to conduct the research and the compiled report will be uploaded to the following Notion page below when finished.\n
\nPlease click the \"Done\" button to complete the form.\n
\n
"
},
"formFields": {
"values": [
{
"html": "=\n\n
\n
![](\"https://res.cloudinary.com/daglih2g8/image/upload/f_auto,q_auto/v1/n8n-workflows/cajjymprexcoesu4gg9g\")
\n
\n
\n
{{ $json.name }}
\n
\n {{ $json.property_description }}\n
\n
\n
\nAnswer the following clarification questions to assist the DeepResearcher better under the research topic.\n
\n
\n\nTotal {{ $('Feedback to Items').all().length }} questions.\n
"
},
"formFields": {
"values": [
{
"fieldType": "textarea",
"fieldLabel": "={{ $json[\"output.questions\"] }}",
"placeholder": "=",
"requiredField": true
}
]
}
},
"typeVersion": 1
},
{
"id": "e07d8c3e-8bcd-4393-9892-f825433ab58d",
"name": "For Each Question...",
"type": "n8n-nodes-base.splitInBatches",
"position": [
-540,
-460
],
"parameters": {
"options": {}
},
"typeVersion": 3
},
{
"id": "e8d26351-52f4-40a6-ba5b-fb6bc816b734",
"name": "DeepResearch Subworkflow",
"type": "n8n-nodes-base.executeWorkflowTrigger",
"position": [
-1880,
820
],
"parameters": {
"workflowInputs": {
"values": [
{
"name": "requestId",
"type": "any"
},
{
"name": "jobType"
},
{
"name": "data",
"type": "object"
}
]
}
},
"typeVersion": 1.1
},
{
"id": "25a8055a-27aa-414f-856b-25a2e2f31974",
"name": "Sticky Note",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1140,
-680
],
"parameters": {
"color": 7,
"width": 1000,
"height": 560,
"content": "## 2. Ask Clarifying Questions\n[Read more about form nodes](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.form/)\n\nTo handle the clarification questions generated by the LLM, I used the same technique found in my \"AI Interviewer\" template ([link](https://n8n.io/workflows/2566-conversational-interviews-with-ai-agents-and-n8n-forms/)).\nThis involves a looping of dynamically generated forms to collect answers from the user."
},
"typeVersion": 1
},
{
"id": "68398b92-eb35-48bf-885e-540074531cc4",
"name": "Clarifying Questions",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
-1040,
-460
],
"parameters": {
"text": "=Given the following query from the user, ask some follow up questions to clarify the research direction. Return a maximum of 3 questions, but feel free to return less if the original query is clear: {{ $json.prompt }}`",
"messages": {
"messageValues": [
{
"type": "HumanMessagePromptTemplate",
"message": "=You are an expert researcher. Today is {{ $now.toLocaleString() }}. Follow these instructions when responding:\n - You may be asked to research subjects that is after your knowledge cutoff, assume the user is right when presented with news.\n - The user is a highly experienced analyst, no need to simplify it, be as detailed as possible and make sure your response is correct.\n - Be highly organized.\n - Suggest solutions that I didn't think about.\n - Be proactive and anticipate my needs.\n - Treat me as an expert in all subject matter.\n - Mistakes erode my trust, so be accurate and thorough.\n - Provide detailed explanations, I'm comfortable with lots of detail.\n - Value good arguments over authorities, the source is irrelevant.\n - Consider new technologies and contrarian ideas, not just the conventional wisdom.\n - You may use high levels of speculation or prediction, just flag it for me."
}
]
},
"promptType": "define",
"hasOutputParser": true
},
"typeVersion": 1.5
},
{
"id": "65c4c293-67b8-4e64-af04-16e45e97c09a",
"name": "Sticky Note1",
"type": "n8n-nodes-base.stickyNote",
"position": [
-660,
-60
],
"parameters": {
"color": 7,
"width": 1360,
"height": 640,
"content": "## 6. Perform DeepSearch Loop\n[Learn more about the Looping in n8n](https://docs.n8n.io/flow-logic/looping/#creating-loops)\n\nThe key of the Deep Research flow is its extensive data collection capability. In this implementation, this capability is represented by a recursive web search & scrape loop which starts with the original query and extended by AI-generated subqueries. How many subqueries to generate are determined the depth and breadth parameters specified.\n\n\"Learnings\" are generated for each subquery and accumulate on each iteration of the loop. When the loop finishes when depth limit is reached, all learnings are collected and it's these learnings are what we use to generate the report."
},
"typeVersion": 1
},
{
"id": "43a5d93d-cae2-43ec-b9ae-b15d6b11b932",
"name": "End Form",
"type": "n8n-nodes-base.form",
"position": [
960,
-420
],
"webhookId": "7b531f5d-942f-4c49-ac55-8ee480889600",
"parameters": {
"options": {},
"operation": "completion",
"completionTitle": "=Thank you for using DeepResearcher.",
"completionMessage": "=You may now close this window."
},
"typeVersion": 1
},
{
"id": "9a824011-e76f-433f-8735-44b358f4ff7d",
"name": "Initiate DeepResearch",
"type": "n8n-nodes-base.executeWorkflow",
"position": [
600,
-420
],
"parameters": {
"mode": "each",
"options": {
"waitForSubWorkflow": false
},
"workflowId": {
"__rl": true,
"mode": "id",
"value": "={{ $workflow.id }}"
},
"workflowInputs": {
"value": {
"data": "={{\n{\n \"query\": $('Get Initial Query').first().json.query,\n \"learnings\": [],\n \"depth\": $('Set Variables').first().json.depth,\n \"breadth\": $('Set Variables').first().json.breadth,\n}\n}}",
"jobType": "deepresearch_initiate",
"requestId": "={{ $('Set Variables').first().json.request_id }}"
},
"schema": [
{
"id": "requestId",
"display": true,
"removed": false,
"required": false,
"displayName": "requestId",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "jobType",
"type": "string",
"display": true,
"removed": false,
"required": false,
"displayName": "jobType",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "data",
"type": "object",
"display": true,
"removed": false,
"required": false,
"displayName": "data",
"defaultMatch": false,
"canBeUsedToMatch": true
}
],
"mappingMode": "defineBelow",
"matchingColumns": [],
"attemptToConvertTypes": false,
"convertFieldsToString": true
}
},
"typeVersion": 1.2
},
{
"id": "c48ee4cd-bac1-4405-bb4c-5614e5eb25a0",
"name": "Page Contents",
"type": "n8n-nodes-base.httpRequest",
"onError": "continueRegularOutput",
"position": [
560,
980
],
"parameters": {
"url": "https://api.apify.com/v2/acts/apify~web-scraper/run-sync-get-dataset-items",
"options": {},
"jsonBody": "={\n \"startUrls\": {{ [{ url: $json.url, method: 'GET' }].toJsonString() }},\n \"breakpointLocation\": \"NONE\",\n \"browserLog\": false,\n \"closeCookieModals\": false,\n \"debugLog\": false,\n \"downloadCss\": false,\n \"downloadMedia\": false,\n \"excludes\": [\n {\n \"glob\": \"/**/*.{png,jpg,jpeg,pdf}\"\n }\n ],\n \"headless\": true,\n \"ignoreCorsAndCsp\": false,\n \"ignoreSslErrors\": false,\n \"injectJQuery\": true,\n \"keepUrlFragments\": false,\n \"linkSelector\": \"\",\n \"maxCrawlingDepth\": 1,\n \"maxPagesPerCrawl\": 1,\n \"maxRequestRetries\": 1,\n \"maxResultsPerCrawl\": 1,\n \"pageFunction\": \"// The function accepts a single argument: the \\\"context\\\" object.\\n// For a complete list of its properties and functions,\\n// see https://apify.com/apify/web-scraper#page-function \\nasync function pageFunction(context) {\\n\\n await new Promise(res => { setTimeout(res, 6000) });\\n // This statement works as a breakpoint when you're trying to debug your code. Works only with Run mode: DEVELOPMENT!\\n // debugger; \\n\\n // jQuery is handy for finding DOM elements and extracting data from them.\\n // To use it, make sure to enable the \\\"Inject jQuery\\\" option.\\n const $ = context.jQuery;\\n const title = $('title').first().text();\\n\\n // Clone the body to avoid modifying the original content\\n const bodyClone = $('body').clone();\\n bodyClone.find('iframe, img, script, style, object, embed, noscript, svg, video, audio').remove();\\n const body = bodyClone.html();\\n\\n // Return an object with the data extracted from the page.\\n // It will be stored to the resulting dataset.\\n return {\\n url: context.request.url,\\n title,\\n body\\n };\\n}\",\n \"postNavigationHooks\": \"// We need to return array of (possibly async) functions here.\\n// The functions accept a single argument: the \\\"crawlingContext\\\" object.\\n[\\n async (crawlingContext) => {\\n // ...\\n },\\n]\",\n \"preNavigationHooks\": \"// We need to return array of (possibly async) functions here.\\n// The functions accept two arguments: the \\\"crawlingContext\\\" object\\n// and \\\"gotoOptions\\\".\\n[\\n async (crawlingContext, gotoOptions) => {\\n // ...\\n },\\n]\\n\",\n \"proxyConfiguration\": {\n \"useApifyProxy\": true\n },\n \"runMode\": \"PRODUCTION\",\n \"useChrome\": false,\n \"waitUntil\": [\n \"domcontentloaded\"\n ],\n \"globs\": [],\n \"pseudoUrls\": [],\n \"proxyRotation\": \"RECOMMENDED\",\n \"maxConcurrency\": 50,\n \"pageLoadTimeoutSecs\": 60,\n \"pageFunctionTimeoutSecs\": 60,\n \"maxScrollHeightPixels\": 5000,\n \"customData\": {}\n}",
"sendBody": true,
"sendQuery": true,
"specifyBody": "json",
"authentication": "genericCredentialType",
"genericAuthType": "httpQueryAuth",
"queryParameters": {
"parameters": [
{
"name": "memory",
"value": "2048"
},
{
"name": "timeout",
"value": "90"
}
]
}
},
"credentials": {
"httpQueryAuth": {
"id": "cO2w8RDNOZg8DRa8",
"name": "Apify API"
}
},
"typeVersion": 4.2
},
{
"id": "dc9f85ff-7565-4c29-981a-5ef65bba6ca3",
"name": "Execution Data",
"type": "n8n-nodes-base.executionData",
"position": [
-1700,
820
],
"parameters": {
"dataToSave": {
"values": [
{
"key": "requestId",
"value": "={{ $json.requestId }}"
},
{
"key": "=jobType",
"value": "={{ $json.jobType }}"
}
]
}
},
"typeVersion": 1
},
{
"id": "26b33429-6d61-4758-9c76-3e998dd31fa4",
"name": "JobType Router",
"type": "n8n-nodes-base.switch",
"position": [
-1520,
820
],
"parameters": {
"rules": {
"values": [
{
"outputKey": "initiate",
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"operator": {
"type": "string",
"operation": "equals"
},
"leftValue": "={{ $json.jobType }}",
"rightValue": "deepresearch_initiate"
}
]
},
"renameOutput": true
},
{
"outputKey": "learnings",
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "ecbfa54d-fc97-48c5-8d3d-f0538b8d727b",
"operator": {
"name": "filter.operator.equals",
"type": "string",
"operation": "equals"
},
"leftValue": "={{ $json.jobType }}",
"rightValue": "deepresearch_learnings"
}
]
},
"renameOutput": true
},
{
"outputKey": "report",
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "392f9a98-ec22-4e57-9c8e-0e1ed6b7dafa",
"operator": {
"name": "filter.operator.equals",
"type": "string",
"operation": "equals"
},
"leftValue": "={{ $json.jobType }}",
"rightValue": "deepresearch_report"
}
]
},
"renameOutput": true
}
]
},
"options": {}
},
"typeVersion": 3.2
},
{
"id": "a9637952-7c09-40ae-96ec-bdf0fc63d94e",
"name": "Valid Pages",
"type": "n8n-nodes-base.filter",
"position": [
720,
980
],
"parameters": {
"options": {},
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "f44691e4-f753-47b0-b66a-068a723b6beb",
"operator": {
"type": "boolean",
"operation": "false",
"singleValue": true
},
"leftValue": "={{ $json['#error'] }}",
"rightValue": ""
},
{
"id": "8e05df2b-0d4a-47da-9aab-da7e8907cbca",
"operator": {
"type": "string",
"operation": "notEmpty",
"singleValue": true
},
"leftValue": "={{ $json.body }}",
"rightValue": ""
}
]
}
},
"typeVersion": 2.2,
"alwaysOutputData": true
},
{
"id": "204cfca2-05bb-46dd-ba96-b41866ed2cfe",
"name": "OpenAI Chat Model4",
"type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
"position": [
-20,
-280
],
"parameters": {
"model": {
"__rl": true,
"mode": "id",
"value": "o3-mini"
},
"options": {}
},
"credentials": {
"openAiApi": {
"id": "8gccIjcuf3gvaoEr",
"name": "OpenAi account"
}
},
"typeVersion": 1.2
},
{
"id": "45bc6261-35c8-4994-bb88-ed7a0f022767",
"name": "Get Initial Query",
"type": "n8n-nodes-base.set",
"position": [
-360,
-540
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "14b77741-c3c3-4bd2-be6e-37bd09fcea2b",
"name": "query",
"type": "string",
"value": "=Initial query: {{ $('Set Variables').first().json.prompt }}\nFollow-up Questions and Answers:\n{{\n$input.all()\n .map(item => {\n const q = Object.keys(item.json)[0];\n const a = item.json[q];\n return `question: ${q}\\nanswer: ${a}`;\n })\n .join('\\n')\n}}"
}
]
}
},
"executeOnce": true,
"typeVersion": 3.4
},
{
"id": "26d26e54-ee9b-4714-ae27-4f033dc825d3",
"name": "Structured Output Parser4",
"type": "@n8n/n8n-nodes-langchain.outputParserStructured",
"position": [
160,
-280
],
"parameters": {
"schemaType": "manual",
"inputSchema": "{\n \"type\": \"object\",\n \"properties\": {\n \"title\": {\n \"type\": \"string\",\n \"description\":\" A short title summarising the research topic\"\n },\n \"description\": {\n \"type\": \"string\",\n \"description\": \"A short description to summarise the research topic\"\n }\n }\n}"
},
"typeVersion": 1.2
},
{
"id": "3842bc1d-d5f9-4879-bc06-db20fed3f55d",
"name": "Create Row",
"type": "n8n-nodes-base.notion",
"position": [
300,
-420
],
"parameters": {
"title": "={{ $json.output.title }}",
"options": {},
"resource": "databasePage",
"databaseId": {
"__rl": true,
"mode": "list",
"value": "19486dd6-0c0c-80da-9cb7-eb1468ea9afd",
"cachedResultUrl": "https://www.notion.so/19486dd60c0c80da9cb7eb1468ea9afd",
"cachedResultName": "n8n DeepResearch"
},
"propertiesUi": {
"propertyValues": [
{
"key": "Description|rich_text",
"textContent": "={{ $json.output.description }}"
},
{
"key": "Status|status",
"statusValue": "Not started"
},
{
"key": "Request ID|rich_text",
"textContent": "={{ $('Set Variables').first().json.request_id }}"
},
{
"key": "Name|title",
"title": "={{ $json.output.title }}"
}
]
}
},
"credentials": {
"notionApi": {
"id": "iHBHe7ypzz4mZExM",
"name": "Notion account"
}
},
"typeVersion": 2.2
},
{
"id": "bfe98996-6ed5-4f60-afdd-a947a6fa6e36",
"name": "Report Page Generator",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
-20,
-420
],
"parameters": {
"text": "=Create a suitable title for the research report which will be created from the user's query.\n{{ $json.query }}",
"promptType": "define",
"hasOutputParser": true
},
"typeVersion": 1.5
},
{
"id": "ff05add8-94b0-4495-8f4e-3e8a10c556af",
"name": "Sticky Note2",
"type": "n8n-nodes-base.stickyNote",
"position": [
-120,
-680
],
"parameters": {
"color": 7,
"width": 600,
"height": 560,
"content": "## 3. Create Empty Report Page in Notion\n[Read more about the Notion node](https://docs.n8n.io/integrations/builtin/app-nodes/n8n-nodes-base.notion/)\n\nSome thought was given where to upload the final report and Notion was selected due to familiarity. This can be easily changed to whatever wiki tools you prefer.\n\nIf you're following along however, here's the Notion database you need to replicate - [Jim's n8n DeepResearcher Database](https://jimleuk.notion.site/19486dd60c0c80da9cb7eb1468ea9afd?v=19486dd60c0c805c8e0c000ce8c87acf)."
},
"typeVersion": 1
},
{
"id": "5bc13d62-81e1-4730-b7e6-9e5579dff174",
"name": "Sticky Note3",
"type": "n8n-nodes-base.stickyNote",
"position": [
500,
-680
],
"parameters": {
"color": 7,
"width": 640,
"height": 560,
"content": "## 4. Trigger DeepResearch Asynchronously\n[Learn more about the Execute Trigger node](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.executeworkflow/)\n\nn8n handles asynchronous jobs by spinning them off as separate executions. This basically means the user doesn't have to wait or keep their browser window open for our researcher to do its job.\n\nOnce we initiate the Deepresearcher job, we can close out the onboarding journey for a nice user experience."
},
"typeVersion": 1
},
{
"id": "9fea6403-b2a2-4e67-99a2-b7a2f29a1e96",
"name": "Sticky Note4",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1160,
620
],
"parameters": {
"color": 7,
"width": 620,
"height": 540,
"content": "## 7. Generate Search Queries\n[Learn more about the Basic LLM node](https://docs.n8n.io/integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainllm/)\n\nMuch like a human researcher, the DeepResearcher will rely on web search and content as the preferred source of information. To ensure it can cover a wide range of sources, the AI can first generate relevant research queries of which each can be explored separately."
},
"typeVersion": 1
},
{
"id": "0bccdc54-7570-4bca-93ec-cb140c5bd3a1",
"name": "Is Depth Reached?",
"type": "n8n-nodes-base.if",
"position": [
-40,
180
],
"parameters": {
"options": {},
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "75d18d88-6ba6-43df-bef7-3e8ad99ad8bd",
"operator": {
"type": "boolean",
"operation": "true",
"singleValue": true
},
"leftValue": "={{ $json.should_stop }}",
"rightValue": ""
}
]
}
},
"typeVersion": 2.2
},
{
"id": "819aa5be-b71b-44a7-b062-b2a50209f290",
"name": "URLs to Items",
"type": "n8n-nodes-base.splitOut",
"position": [
400,
980
],
"parameters": {
"options": {},
"fieldToSplitOut": "results"
},
"typeVersion": 1
},
{
"id": "4b8e9936-4b24-4bd4-8fe7-75d58244cb6d",
"name": "Get Research Results",
"type": "n8n-nodes-base.set",
"position": [
160,
180
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "90b3da00-dcd5-4289-bd45-953146a3b0ba",
"name": "all_learnings",
"type": "array",
"value": "={{ $json.all_learnings }}"
},
{
"id": "623dbb3d-83a1-44a9-8ad3-48d92bc42811",
"name": "all_urls",
"type": "array",
"value": "={{ $json.all_urls }}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "d371535a-2946-4ec5-9be6-2ee8e359ac44",
"name": "Get Existing Row",
"type": "n8n-nodes-base.notion",
"position": [
-1040,
180
],
"parameters": {
"limit": 1,
"filters": {
"conditions": [
{
"key": "Request ID|rich_text",
"condition": "equals",
"richTextValue": "={{ $json.requestId.toString() }}"
}
]
},
"options": {},
"resource": "databasePage",
"matchType": "allFilters",
"operation": "getAll",
"databaseId": {
"__rl": true,
"mode": "list",
"value": "19486dd6-0c0c-80da-9cb7-eb1468ea9afd",
"cachedResultUrl": "https://www.notion.so/19486dd60c0c80da9cb7eb1468ea9afd",
"cachedResultName": "n8n DeepResearch"
},
"filterType": "manual"
},
"credentials": {
"notionApi": {
"id": "iHBHe7ypzz4mZExM",
"name": "Notion account"
}
},
"typeVersion": 2.2
},
{
"id": "fea4c30e-1193-494d-8823-dfbec5196a0d",
"name": "Set In-Progress",
"type": "n8n-nodes-base.notion",
"position": [
-840,
180
],
"parameters": {
"pageId": {
"__rl": true,
"mode": "id",
"value": "={{ $json.id }}"
},
"options": {},
"resource": "databasePage",
"operation": "update",
"propertiesUi": {
"propertyValues": [
{
"key": "Status|status",
"statusValue": "In progress"
}
]
}
},
"credentials": {
"notionApi": {
"id": "iHBHe7ypzz4mZExM",
"name": "Notion account"
}
},
"typeVersion": 2.2
},
{
"id": "37954acd-d8cb-4c74-afa8-d8973e017327",
"name": "Set Done",
"type": "n8n-nodes-base.notion",
"position": [
1680,
1600
],
"parameters": {
"pageId": {
"__rl": true,
"mode": "id",
"value": "={{ $('Get Existing Row1').first().json.id }}"
},
"options": {},
"resource": "databasePage",
"operation": "update",
"propertiesUi": {
"propertyValues": [
{
"key": "Status|status",
"statusValue": "Done"
},
{
"key": "Last Updated|date",
"date": "={{ $now.toISO() }}"
}
]
}
},
"credentials": {
"notionApi": {
"id": "iHBHe7ypzz4mZExM",
"name": "Notion account"
}
},
"executeOnce": true,
"typeVersion": 2.2
},
{
"id": "3db97ab8-b934-4567-a92e-92374a363df6",
"name": "Tags to Items",
"type": "n8n-nodes-base.splitOut",
"position": [
-60,
1600
],
"parameters": {
"options": {},
"fieldToSplitOut": "tag"
},
"typeVersion": 1
},
{
"id": "7f468bf7-762a-4818-86f9-54d172bb618a",
"name": "Convert to HTML",
"type": "n8n-nodes-base.markdown",
"position": [
-380,
1600
],
"parameters": {
"mode": "markdownToHtml",
"options": {
"tables": true
},
"markdown": "={{ $json.text }}"
},
"typeVersion": 1
},
{
"id": "97914ee9-0ee8-408b-b2bb-a7193b2d0454",
"name": "HTML to Array",
"type": "n8n-nodes-base.set",
"position": [
-220,
1600
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "851b8a3f-c2d3-41ad-bf60-4e0e667f6c58",
"name": "tag",
"type": "array",
"value": "={{ $json.data.match(/||<[^>]+>[^<]*<\\/[^>]+>/g) }}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "6ce79f16-51e3-4192-8103-738222be558b",
"name": "Notion Block Generator",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
100,
1600
],
"parameters": {
"text": "={{ $json.tag.trim() }}",
"messages": {
"messageValues": [
{
"message": "=Convert the following html into its equivalent Notion Block as per Notion's API schema.\n* Ensure the content is always included and remains the same.\n* Return only a json response.\n* Generate child-level blocks. Should not define \"parent\" or \"children\" property.\n* Strongly prefer headings, paragraphs, tables and lists type blocks.\n* available headings are heading_1, heading_2 and heading_3 - h4,h5,h6 should use heading_3 type instead. ensure headings use the rich text definition.\n* ensure lists blocks include all list items.\n\n## Examples\n\n1. headings\n```\nReferences
\n```\nwould convert to \n```\n{\"object\": \"block\", \"type\": \"heading_3\", \"heading_3\": { \"rich_text\": [{\"type\": \"text\",\"text\": {\"content\": \"References\"}}]}}\n```\n\n2. lists\n```\n\n```\nwould convert to\n```\n[\n{\n \"object\": \"block\",\n \"type\": \"bulleted_list_item\",\n \"bulleted_list_item\": {\"rich_text\": [{\"type\": \"text\",\"text\": {\"content\": \"hello\"}}]}\n},\n{\n \"object\": \"block\",\n \"type\": \"bulleted_list_item\",\n \"bulleted_list_item\": {\"rich_text\": [{\"type\": \"text\",\"text\": {\"content\": \"world\"}}]}\n}\n]\n```\n\n3. tables\n```\n\n \n | Technology | Potential Impact |
|---|
\n \n \n \n | 5G Connectivity | Enables faster data speeds and advanced apps | \n
\n \n
\n```\nwould convert to\n```\n{\n \"object\": \"block\",\n \"type\": \"table\",\n \"table\": {\n \"table_width\": 2,\n \"has_column_header\": true,\n \"has_row_header\": false,\n \"children\": [\n {\n \"object\": \"block\",\n \"type\": \"table_row\",\n \"table_row\": {\n \"cells\": [\n [\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \"Technology\",\n \"link\": null\n }\n },\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \"Potential Impact\",\n \"link\": null\n }\n }\n ],\n [\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \"5G Connectivity\",\n \"link\": null\n }\n },\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \"Enables faster data speeds and advanced apps\",\n \"link\": null\n }\n }\n ]\n ]\n }\n }\n ]\n }\n}\n```\n4. anchor links\nSince Notion doesn't support anchor links, just convert them to rich text blocks instead.\n```\nModule 0: Pre-Course Setup and Learning Principles\n```\nconverts to\n```\n{\n \"object\": \"block\",\n \"type\": \"paragraph\",\n \"paragraph\": {\n \"rich_text\": [\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \"Module 0: Pre-Course Setup and Learning Principles\"\n }\n }\n ]\n }\n}\n```\n5. Invalid html parts\nWhen the html is not syntax valid eg. orphaned closing tags, then just skip the conversion and use an empty rich text block.\n```\n\\n\n```\ncan be substituted with\n```\n{\n \"object\": \"block\",\n \"type\": \"paragraph\",\n \"paragraph\": {\n \"rich_text\": [\n {\n \"type\": \"text\",\n \"text\": {\n \"content\": \" \"\n }\n }\n ]\n }\n}\n```"
}
]
},
"promptType": "define"
},
"typeVersion": 1.5
},
{
"id": "e3eeb9f0-7407-41f9-a814-def6c26b2ee1",
"name": "Google Gemini Chat Model",
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"position": [
80,
1760
],
"parameters": {
"options": {},
"modelName": "models/gemini-2.0-flash"
},
"credentials": {
"googlePalmApi": {
"id": "dSxo6ns5wn658r8N",
"name": "Google Gemini(PaLM) Api account"
}
},
"typeVersion": 1
},
{
"id": "5b0aeaca-dce5-4afd-8a8a-0ef2c18b6f06",
"name": "Parse JSON blocks",
"type": "n8n-nodes-base.set",
"onError": "continueRegularOutput",
"position": [
420,
1600
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "73fcb8a0-2672-4bd5-86de-8075e1e02baf",
"name": "=block",
"type": "array",
"value": "={{\n(function(){\n const block = $json.text\n .replace('```json', '')\n .replace('```', '')\n .trim()\n .parseJson();\n if (Array.isArray(block)) return block;\n if (block.type.startsWith('heading_')) {\n const prev = Number(block.type.split('_')[1]);\n const next = Math.max(1, prev - 1);\n if (next !== prev) {\n block.type = `heading_${next}`;\n block[`heading_${next}`] = Object.assign({}, block[`heading_${prev}`]);\n block[`heading_${prev}`] = undefined;\n }\n }\n return [block];\n})()\n}}"
}
]
}
},
"executeOnce": false,
"typeVersion": 3.4
},
{
"id": "e2a5a5bc-a3c8-42c5-9419-74ce3525f599",
"name": "Upload to Notion Page",
"type": "n8n-nodes-base.httpRequest",
"onError": "continueRegularOutput",
"maxTries": 2,
"position": [
1680,
1760
],
"parameters": {
"url": "=https://api.notion.com/v1/blocks/{{ $('Get Existing Row1').first().json.id }}/children",
"method": "PATCH",
"options": {
"timeout": "={{ 1000 * 60 }}"
},
"jsonBody": "={{\n{\n \"children\": $json.block\n}\n}}",
"sendBody": true,
"sendHeaders": true,
"specifyBody": "json",
"authentication": "predefinedCredentialType",
"headerParameters": {
"parameters": [
{
"name": "Notion-Version",
"value": "2022-06-28"
}
]
},
"nodeCredentialType": "notionApi"
},
"credentials": {
"notionApi": {
"id": "iHBHe7ypzz4mZExM",
"name": "Notion account"
}
},
"retryOnFail": true,
"typeVersion": 4.2,
"waitBetweenTries": 3000
},
{
"id": "9f18b2a5-ba74-40fc-8e35-a93ecd13507a",
"name": "Sticky Note5",
"type": "n8n-nodes-base.stickyNote",
"position": [
-520,
620
],
"parameters": {
"color": 7,
"width": 1740,
"height": 740,
"content": "## 8. Web Search and Extracting Web Page Contents using [APIFY.com](https://www.apify.com?fpr=414q6)\n[Read more about the HTTP Request node](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.httprequest/)\n\nHere is where I deviated a little from the reference implementation. I opted not to use Firecrawl.ai due to (1) high cost of the service and (2) a regular non-ai crawler would work just as well and probably quicker.\nInstead I'm using [APIFY.com](https://www.apify.com?fpr=414q6) which is a more performant, cost-effective and reliable web scraper service. If you don't want to use Apify, feel free to swap this out with your preferred service.\n\nThis step is the most exciting in terms of improvements and optimisations eg. mix in internal data sources! Add in Perplexity.ai or Jina.ai! Possibilities are endless."
},
"typeVersion": 1
},
{
"id": "84c34a2a-d8bb-4e62-a5ea-df0a142aa2b4",
"name": "Sticky Note6",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1140,
60
],
"parameters": {
"color": 7,
"width": 460,
"height": 360,
"content": "## 5. Set Report to In-Progress\n[Read more about the Notion node](https://docs.n8n.io/integrations/builtin/app-nodes/n8n-nodes-base.notion/)"
},
"typeVersion": 1
},
{
"id": "bd022636-873d-4aca-8929-c189f8596cc1",
"name": "Sticky Note7",
"type": "n8n-nodes-base.stickyNote",
"position": [
1240,
700
],
"parameters": {
"color": 7,
"width": 780,
"height": 660,
"content": "## 9. Compile Learnings with Reasoning Model\n[Read more about the Basic LLM node](https://docs.n8n.io/integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainllm/)\n\nWith our gathered sources, it's now just a case of giving it to our LLM to compile a list of \"learnings\" from them. For our DeepResearcher, we'll use OpenAI's o3-mini which is the latest reasoning model at time of writing. Reasoning perform better than regular chat models due their chain-of-thought or \"thinking\" process that they perform.\n\nThe \"Learnings\" are then combined with the generated research goal to complete one loop."
},
"typeVersion": 1
},
{
"id": "b2f6e51d-cbe6-4459-9515-679f79063926",
"name": "Get Existing Row1",
"type": "n8n-nodes-base.notion",
"position": [
-1020,
1600
],
"parameters": {
"limit": 1,
"filters": {
"conditions": [
{
"key": "Request ID|rich_text",
"condition": "equals",
"richTextValue": "={{ $json.requestId.toString() }}"
}
]
},
"options": {},
"resource": "databasePage",
"matchType": "allFilters",
"operation": "getAll",
"databaseId": {
"__rl": true,
"mode": "list",
"value": "19486dd6-0c0c-80da-9cb7-eb1468ea9afd",
"cachedResultUrl": "https://www.notion.so/19486dd60c0c80da9cb7eb1468ea9afd",
"cachedResultName": "n8n DeepResearch"
},
"filterType": "manual"
},
"credentials": {
"notionApi": {
"id": "iHBHe7ypzz4mZExM",
"name": "Notion account"
}
},
"typeVersion": 2.2
},
{
"id": "7b315060-7e40-410c-ac9d-ef22acbb175a",
"name": "Sticky Note8",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1140,
1400
],
"parameters": {
"color": 7,
"width": 660,
"height": 540,
"content": "## 10. Generate DeepSearch Report using Learnings\n[Read more about the Basic LLM node](https://docs.n8n.io/integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainllm/)\n\nFinally! After all learnings have been gathered - which may have taken up to an hour or more on the higher settings! - they are given to our LLM to generate the final research report in markdown format. Technically, the DeepResearch ends here but for this template, we need to push the output to Notion. If you're not using Notion, feel free to ignore the last few steps."
},
"typeVersion": 1
},
{
"id": "01aaf2a8-c145-4f39-aa52-f40fc28f8767",
"name": "Sticky Note9",
"type": "n8n-nodes-base.stickyNote",
"position": [
-460,
1400
],
"parameters": {
"color": 7,
"width": 1060,
"height": 540,
"content": "## 11. Reformat Report as Notion Blocks\n[Learn more about the Markdown node](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.markdown/)\n\nTo write our report to our Notion page, we'll have to convert it to Notion \"blocks\" - these are specialised json objects which are required by the Notion API. There are quite a number of ways to do this conversion not involving the use of AI but for kicks, I decided to do so anyway. In this step, we first convert to HTML as it allows us to split the report semantically and makes for easier parsing for the LLM."
},
"typeVersion": 1
},
{
"id": "700a6f44-86bf-4aab-8a42-23bf6843f681",
"name": "Sticky Note10",
"type": "n8n-nodes-base.stickyNote",
"position": [
1220,
1400
],
"parameters": {
"color": 7,
"width": 800,
"height": 580,
"content": "## 13. Update Report in Notion\n[Read more about the HTTP request node](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.httprequest/)\n\nIn this step, we can use the Notion API to add the blocks to our page sequentially. A loop is used due to the unstable Notion API - the loop allows retries for blocks that require it."
},
"typeVersion": 1
},
{
"id": "f8536052-c851-42ec-aaf3-5fc876570f6d",
"name": "Sticky Note11",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1840,
-680
],
"parameters": {
"color": 7,
"width": 680,
"height": 560,
"content": "## 1. Let's Research!\n[Learn more about the form trigger node](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.formtrigger)\n\nn8n forms are a really nice way to get our frontend up and running quickly and compared to chat, offers a superior user interface for user input. I've gone perhaps a little extra with the custom html fields but I do enjoy adding a little customisation now and then."
},
"typeVersion": 1
},
{
"id": "58aa92a2-e1fd-497d-a27e-40b733189bab",
"name": "DeepResearch Report",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
-860,
1600
],
"parameters": {
"text": "=You are are an expert and insightful researcher.\n* Given the following prompt from the user, write a final report on the topic using the learnings from research.\n* Make it as as detailed as possible, aim for 3 or more pages, include ALL the learnings from research.\n* Format the report in markdown. Use headings, lists and tables only and where appropriate.\n\n{{ $('JobType Router').first().json.data.query }}\n\nHere are all the learnings from previous research:\n\n\n{{\n$('JobType Router').first().json.data\n .all_learnings\n .map(item => `${item}`) \n .join('\\n')\n}}\n",
"promptType": "define"
},
"typeVersion": 1.5
},
{
"id": "0656be83-d510-46f1-aeeb-f62a69aa3cf2",
"name": "DeepResearch Learnings",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
1500,
980
],
"parameters": {
"text": "=Given the following contents from a SERP search for the query {{ $('Item Ref').first().json.query }}, generate a list of learnings from the contents. Return a maximum of 3 learnings, but feel free to return less if the contents are clear. Make sure each learning is unique and not similar to each other. The learnings should be concise and to the point, as detailed and infromation dense as possible. Make sure to include any entities like people, places, companies, products, things, etc in the learnings, as well as any exact metrics, numbers, or dates. The learnings will be used to research the topic further.\n\n\n{{\n$('Convert to Markdown')\n .all()\n .map(item =>`\\n${item.json.markdown.substr(0, 25_000)}\\n`)\n .join('\\n')\n}}\n",
"messages": {
"messageValues": [
{
"type": "HumanMessagePromptTemplate",
"message": "=You are an expert researcher. Today is {{ $now.toLocaleString() }}. Follow these instructions when responding:\n - You may be asked to research subjects that is after your knowledge cutoff, assume the user is right when presented with news.\n - The user is a highly experienced analyst, no need to simplify it, be as detailed as possible and make sure your response is correct.\n - Be highly organized.\n - Suggest solutions that I didn't think about.\n - Be proactive and anticipate my needs.\n - Treat me as an expert in all subject matter.\n - Mistakes erode my trust, so be accurate and thorough.\n - Provide detailed explanations, I'm comfortable with lots of detail.\n - Value good arguments over authorities, the source is irrelevant.\n - Consider new technologies and contrarian ideas, not just the conventional wisdom.\n - You may use high levels of speculation or prediction, just flag it for me."
}
]
},
"promptType": "define",
"hasOutputParser": true
},
"executeOnce": true,
"typeVersion": 1.5
},
{
"id": "9296a787-3226-44fe-8118-f84dda8e5167",
"name": "Generate Report",
"type": "n8n-nodes-base.executeWorkflow",
"position": [
480,
180
],
"parameters": {
"options": {
"waitForSubWorkflow": false
},
"workflowId": {
"__rl": true,
"mode": "id",
"value": "={{ $workflow.id }}"
},
"workflowInputs": {
"value": {
"data": "={{\n{\n ...Object.assign({}, $json),\n query: $('JobType Router').first().json.data.query\n}\n}}",
"jobType": "deepresearch_report",
"requestId": "={{ $('JobType Router').first().json.requestId }}"
},
"schema": [
{
"id": "requestId",
"display": true,
"required": false,
"displayName": "requestId",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "jobType",
"type": "string",
"display": true,
"required": false,
"displayName": "jobType",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "data",
"type": "object",
"display": true,
"required": false,
"displayName": "data",
"defaultMatch": false,
"canBeUsedToMatch": true
}
],
"mappingMode": "defineBelow",
"matchingColumns": [],
"attemptToConvertTypes": false,
"convertFieldsToString": true
}
},
"typeVersion": 1.2
},
{
"id": "471f9b9f-f331-4652-95de-1ec7136ea692",
"name": "Generate Learnings",
"type": "n8n-nodes-base.executeWorkflow",
"position": [
-380,
180
],
"parameters": {
"mode": "each",
"options": {
"waitForSubWorkflow": true
},
"workflowId": {
"__rl": true,
"mode": "id",
"value": "={{ $workflow.id }}"
},
"workflowInputs": {
"value": {
"data": "={{ $json }}",
"jobType": "deepresearch_learnings",
"requestId": "={{ $('JobType Router').first().json.requestId }}"
},
"schema": [
{
"id": "requestId",
"display": true,
"removed": false,
"required": false,
"displayName": "requestId",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "jobType",
"type": "string",
"display": true,
"removed": false,
"required": false,
"displayName": "jobType",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "data",
"type": "object",
"display": true,
"removed": false,
"required": false,
"displayName": "data",
"defaultMatch": false,
"canBeUsedToMatch": true
}
],
"mappingMode": "defineBelow",
"matchingColumns": [],
"attemptToConvertTypes": false,
"convertFieldsToString": true
}
},
"typeVersion": 1.2
},
{
"id": "2e2fa1e6-9d5c-46ff-985c-58ada1139837",
"name": "Confirmation",
"type": "n8n-nodes-base.form",
"position": [
780,
-420
],
"webhookId": "cf41a176-5d30-4274-955e-b0d5b483d37f",
"parameters": {
"options": {
"formTitle": "DeepResearcher",
"buttonLabel": "Done",
"formDescription": "=\n\nYour Report Is On Its Way!\n
\nDeepResearcher will now work independently to conduct the research and the compiled report will be uploaded to the following Notion page below when finished.\n
\nPlease click the \"Done\" button to complete the form.\n
\n
"
},
"formFields": {
"values": [
{
"html": "=\n\n
\n
\n
\n
\n
{{ $json.name }}
\n
\n {{ $json.property_description }}\n
\n
\n