January 22, 2026

Brave Search to Google Sheets, ranked research ready

Lisa Granqvist Partner Workflow Automation Expert

Get a free AI assessment → ⬇️ Use template

Research work breaks in a quiet way. You open ten tabs, copy a few links, lose the best source, then paste a messy list into a sheet you’ll “clean up later” (you won’t).

This hits content strategists building briefs, but growth marketers and agency leads feel it too. Brave Sheets automation turns a raw question into a ranked top 10 list in Google Sheets, so you stop doing the same search three different times.

Below, you’ll see how the workflow turns a webhook request into refined search, semantic re-ranking, and a clean, reusable output you can drop straight into planning.

How This Automation Works

See how this solves the problem:

n8n Workflow Template: Brave Search to Google Sheets, ranked research ready

Click to explore

flowchart LR

    subgraph sg0["Auto-fixing Output P Flow"]
        direction LR
        n0@{ icon: "mdi:cog", form: "rounded", label: "Date & Time", pos: "b", h: 48 }
        n1["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/webhook.dark.svg' width='40' height='40' /></div><br/>Webhook"]
        n2@{ icon: "mdi:robot", form: "rounded", label: "Auto-fixing Output Parser6", pos: "b", h: 48 }
        n3@{ icon: "mdi:robot", form: "rounded", label: "Auto-fixing Output Parser", pos: "b", h: 48 }
        n4@{ icon: "mdi:robot", form: "rounded", label: "Structured Output Parser1", pos: "b", h: 48 }
        n5["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/code.svg' width='40' height='40' /></div><br/>Query-1 Combined"]
        n6["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/webhook.dark.svg' width='40' height='40' /></div><br/>Respond to Webhook"]
        n7@{ icon: "mdi:robot", form: "rounded", label: "Semantic Search - Result Re-..", pos: "b", h: 48 }
        n8["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/httprequest.dark.svg' width='40' height='40' /></div><br/>Query"]
        n10@{ icon: "mdi:robot", form: "rounded", label: "Semantic Search -Query Maker", pos: "b", h: 48 }
        n13@{ icon: "mdi:robot", form: "rounded", label: "Structured Output Parser2", pos: "b", h: 48 }
        n14@{ icon: "mdi:brain", form: "rounded", label: "Parser Model", pos: "b", h: 48 }
        n15@{ icon: "mdi:brain", form: "rounded", label: "Agent Model", pos: "b", h: 48 }
        n8 --> n5
        n1 --> n0
        n15 -.-> n7
        n15 -.-> n10
        n0 --> n10
        n14 -.-> n2
        n14 -.-> n3
        n5 --> n7
        n3 -.-> n10
        n4 -.-> n3
        n13 -.-> n2
        n2 -.-> n7
        n10 --> n8
        n7 --> n6
    end

    subgraph sg1["Flow 2"]
        direction LR
        n11@{ icon: "mdi:brain", form: "rounded", label: "Anthropic Chat Model", pos: "b", h: 48 }
    end

    subgraph sg2["Flow 3"]
        direction LR
        n12@{ icon: "mdi:brain", form: "rounded", label: "OpenAI Chat Model", pos: "b", h: 48 }
    end

    subgraph sg3["Flow 4"]
        direction LR
        n9["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/httprequest.dark.svg' width='40' height='40' /></div><br/>Webhook Call"]
    end

    %% Styling
    classDef trigger fill:#e8f5e9,stroke:#388e3c,stroke-width:2px
    classDef ai fill:#e3f2fd,stroke:#1976d2,stroke-width:2px
    classDef aiModel fill:#e8eaf6,stroke:#3f51b5,stroke-width:2px
    classDef decision fill:#fff8e1,stroke:#f9a825,stroke-width:2px
    classDef database fill:#fce4ec,stroke:#c2185b,stroke-width:2px
    classDef api fill:#fff3e0,stroke:#e65100,stroke-width:2px
    classDef code fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px
    classDef disabled stroke-dasharray: 5 5,opacity: 0.5
    class n2,n3,n4,n7,n10,n13 ai
    class n14,n15,n11,n12 aiModel
    class n1,n6,n8,n9 api
    class n5 code
    classDef customIcon fill:none,stroke:none
    class n1,n5,n6,n8,n9 customIcon

The Challenge: Turning messy search into usable research

Most “research” isn’t research. It’s collecting links under time pressure, trying to remember why a source mattered, and hoping you didn’t miss the obvious result buried on page two. The worst part is the repeat work. You search, skim, copy, paste, then someone asks the same question next week and you do it all again. Add a couple of teammates, and now you’re also reconciling different search phrases and different standards for what counts as “good.”

It adds up fast. Here’s where it usually breaks down in real life.

You lose time rewriting queries because the first version was too broad or too vague.
Top results look “fine,” but they are not actually relevant to the brief, so you waste another hour hunting.
Copy-pasting titles, URLs, and snippets into Google Sheets creates typos, duplicates, and half-finished rows.
No consistent ranking method means the “best sources” depend on who searched that day.

The Fix: AI-ranked Brave Search results logged to Sheets

This workflow acts like a research assistant that never gets tired. You send a question into a webhook, and the automation immediately timestamps the request so your output is traceable later. Then Google Gemini rewrites your plain-English question into a tighter, more “expert” query that Brave Search can answer more precisely. Brave Search returns results, and the workflow aggregates them into a consistent structure. Finally, Gemini semantically re-ranks those results against your original intent, so your top 10 is about relevance, not just SEO luck. The output is structured (titles, links, descriptions, and ranks), which makes it easy to log into Google Sheets and reuse for briefs.

The workflow starts with an incoming request (Webhook) and a timestamp (Date & Time). From there, Brave Search is called via HTTP Request, then an AI chain reorders results by meaning, not keywords. At the end, you get a clean top 10 list you can store and share.

What Changes: Before vs. After

What This Eliminates

Impact You’ll See

Manually rewriting search queries when the first results miss the mark.
Skimming 20+ links because ranking is based on generic SEO, not your brief.
Copy-pasting titles, URLs, and snippets into a sheet one row at a time.
Reformatting outputs each time because every researcher structures it differently.

Most teams get about 1–2 hours back per research task.
Your “top 10” stays consistent because relevance is scored the same way.
Briefs become faster to assemble because sources arrive pre-structured.
Less second-guessing, since the reranking matches the original question.
Easy handoffs to clients or teammates without extra explanation.

Real-World Impact

Say you’re building one competitive research brief per day. Manually, a typical cycle looks like 10 minutes refining the query, 30 minutes opening and scanning results, then about 20 minutes copying the best 10 into Google Sheets (roughly an hour total). With this workflow, you send the question once (a minute or two), wait for Brave + Gemini to generate and rerank, and you’re done in about 10 minutes of real effort. That’s close to an hour back, per brief, without lowering quality.

Requirements

n8n instance (try n8n Cloud free)
Self-hosting option if you prefer (Hostinger works well)
Brave Search API for pulling web search results.
Google Gemini to refine queries and rerank results.
Google Sheets to store ranked top 10 results.
Brave Search API key (get it from api.search.brave.com).

Skill level: Intermediate. You’ll paste API keys, test a webhook, and map a few fields into a sheet.

Need help implementing this? Talk to an automation expert (free 15-minute consultation).

The Workflow Flow

A webhook captures the question. You send a search query from another tool (a form, a button, a doc workflow), and n8n receives it instantly through the Incoming Request Trigger.

The query gets upgraded before searching. The workflow adds a timestamp, then uses a Gemini model to turn your request into a sharper “semantic” query that matches the intent behind the words.

Brave Search pulls results, then AI reranks them. An HTTP Request calls the Brave Search API, results are aggregated, and a semantic reordering step sorts them by relevance to the original question. This is the part that usually saves you from page-two regret.

Structured output goes to your systems. The workflow returns a JSON response via webhook, and it can log the same ranked list into Google Sheets (and, if you want, into Excel 365 or Google Drive workflows already running in your stack).

You can easily modify the ranking criteria to prioritize recency, source type, or specific keywords based on your needs. See the full implementation guide below for customization options.

Step-by-Step Implementation Guide

Step 1: Configure the Webhook Trigger

Set up the incoming webhook endpoint that starts the workflow and provides the research question input.

Add Incoming Request Trigger and set Path to 962f1468-c80f-4c0c-8555-a0acf648ede4.
Set Response Mode to responseNode so the workflow returns results via Return Webhook Response.
Connect Incoming Request Trigger to Current Time Stamp to pass the incoming payload into the workflow.

Tip: Use Utility: Webhook Test Call with Research Question set to what is the latest news in global world in politics and economy? to simulate a request during testing.

Step 2: Generate Context and Compose the Search Query

Create a time-aware query using AI, then structure the model output with schema and auto-fix parsing.

In Current Time Stamp, keep default options to generate currentDate.
Connect Current Time Stamp to Search Query Composer.
In Search Query Composer, set Text to the provided prompt and keep Prompt Type as define.
Ensure the expressions remain intact, including {{ $item("0").$node["Current Time Stamp"].json["currentDate"] }} and {{ $item("0").$node["Incoming Request Trigger"].json["query"]["Research Question"] }}.
Attach Schema Parser Alpha → Auto Repair Parser A as the output parser chain for Search Query Composer.
Credential Required: Connect your googlePalmApi credentials on Gemini Agent Model, which is the language model for Search Query Composer.

⚠️ Common Pitfall: Do not add credentials directly to Schema Parser Alpha or Auto Repair Parser A. These are AI sub-nodes and must use credentials on the parent model node (Gemini Agent Model).

Step 3: Call the Web Search API

Use the generated query to call the Brave Search API and retrieve raw results.

Connect Search Query Composer to Primary Web Search Call.
In Primary Web Search Call, set URL to https://api.search.brave.com/res/v1/web/search.
Enable Send Query and set the Query Parameters name to q with value {{ $json.output.final_search_query }}.
Enable Send Headers and set X-Subscription-Token to [CONFIGURE_YOUR_API_KEY].

⚠️ Common Pitfall: The Brave API call will fail if X-Subscription-Token is not replaced with a valid key.

Step 4: Aggregate and Re-Rank Results with AI

Transform the raw results into a single text block, then rank and extract insights using a second AI chain.

Connect Primary Web Search Call to Aggregate Search Results.
In Aggregate Search Results, keep the provided JavaScript Code to build aggregated_text from web.results.
Connect Aggregate Search Results to Semantic Result Reorder.
In Semantic Result Reorder, keep the long-form prompt and make sure it references {{ $json.aggregated_text }} and the research question expression {{ $('Incoming Request Trigger').item.json.query['Research Question'] }}.
Attach Schema Parser Beta → Auto Repair Parser B as the output parser chain for Semantic Result Reorder.
Credential Required: Connect your googlePalmApi credentials on Gemini Agent Model, which is the language model for Semantic Result Reorder.

Tip: Gemini Parsing Model provides parsing support for Auto Repair Parser A and Auto Repair Parser B; ensure its googlePalmApi credential is configured for reliable auto-fixing.

Step 5: Return the Structured Webhook Response

Send the ranked URLs and extracted information back to the webhook caller.

Connect Semantic Result Reorder to Return Webhook Response.
Set Respond With to text.
Set Response Body to the JSON template that maps fields from Semantic Result Reorder, for example: {{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_1']['title'] }}.

Step 6: Review Optional Utility Models

Two additional LLM nodes are present for potential future expansion or alternate routing.

Utility: Claude Chat Model is available with Credential Required: Connect your anthropicApi credentials if you plan to use this node.
Utility: OpenAI Dialogue Model is available with Credential Required: Connect your openAiApi credentials if you plan to use this node.

Tip: These utility models are not connected in the current execution path. You can safely leave them unused, or wire them into future branches.

Step 7: Test and Activate Your Workflow

Validate the end-to-end flow and confirm the webhook returns ranked results before enabling production use.

Click Execute Workflow and trigger Incoming Request Trigger using Utility: Webhook Test Call or a manual POST request.
Confirm that Primary Web Search Call returns results and Aggregate Search Results outputs aggregated_text.
Verify that Semantic Result Reorder outputs a JSON object with ranked URLs and Information_extracted.
Check that Return Webhook Response returns the structured JSON payload to the caller.
When satisfied, toggle the workflow to Active for production use.

🔒

Unlock Full Step-by-Step Guide

Get the complete implementation guide + downloadable template

Watch Out For

Brave Search API credentials can expire or need specific permissions. If things break, check your Brave API dashboard (and the header value inside the HTTP Request node) first.
If you’re using Wait nodes or external rendering, processing times vary. Bump up the wait duration if downstream nodes fail on empty responses.
Default prompts in AI nodes are generic. Add your brand voice early or you’ll be editing outputs forever.

Common Questions

How quickly can I implement this Brave Sheets automation?

About 30 minutes if your API keys are ready.

Can non-technical teams implement this Brave Sheets automation?

Yes. You’ll connect accounts, paste API keys, and test the webhook once.

Is n8n free to use for this Brave Sheets automation workflow?

Yes. n8n has a free self-hosted option and a free trial on n8n Cloud. Cloud plans start at $20/month for higher volume. You’ll also need to factor in Gemini API usage (often a few cents per run) plus whatever Brave Search tier you’re on.

Where can I host n8n to run this automation?

Two options: n8n Cloud (managed, easiest setup) or self-hosting on a VPS. For self-hosting, Hostinger VPS is affordable and handles n8n well. Self-hosting gives you unlimited executions but requires basic server management.

How do I adapt this Brave Sheets automation solution to my specific challenges?

You can swap the Brave Search call for another search provider by changing the HTTP Request node, then keeping the same “aggregate → semantic reorder → output” pattern. Most teams customize the Gemini prompts in the Search Query Composer and Semantic Result Reorder nodes to match their niche (for example: “only include sources from the last 12 months” or “prioritize pricing pages and official docs”). If you want the output in a different structure, adjust the structured parser nodes so your Google Sheets columns stay stable.

Why is my Brave Search connection failing in this workflow?

Usually it’s a missing or expired Brave API key in the HTTP Request headers. It can also be rate limiting on the free tier if you run bursts, so check Brave’s response code and message in the node output. Finally, confirm you’re sending a valid query string into the webhook, because an empty input can look like a “search failure” downstream.

What’s the capacity of this Brave Sheets automation solution?

On n8n Cloud Starter you can run thousands of executions per month; higher plans handle more. If you self-host, there’s no hard execution cap (it depends on your server). In practice, this workflow runs as fast as Brave Search and the Gemini reranking step return responses, so heavy usage is usually limited by API quotas, not n8n itself.

Is this Brave Sheets automation better than using Zapier or Make?

Often, yes, because this is not a simple “search → log” zap. You’re doing query rewriting, aggregation, and semantic reranking, which usually means branching logic and structured parsing. n8n handles that kind of flow without forcing you into expensive task pricing as quickly, and self-hosting is a real option if you want to run a lot of research jobs. Zapier or Make can still work if you only need a lightweight version (for example: one search call and a single sheet row). If you’re torn, Talk to an automation expert and map the cheapest path for your volume.

Once this is in place, research stops being a time sink and starts behaving like an input you can reliably reuse. The workflow handles the repetitive sorting and formatting, and you get to focus on the actual brief.

Need Help Setting This Up?

Our automation experts can build and customize this workflow for your specific needs. Free 15-minute consultation—no commitment required.

Lisa Granqvist

Workflow Automation Expert

Expert in workflow automation and no-code tools.

{
"id": null,
"meta": {
"instanceId": "902e76a8eeb5c6bcda4a8e7ee81223c1f7fa98dcebe1a2aa2b901d40049f0629"
},
"name": "Automated Web Query Re-Ranking System",
"tags": [],
"nodes": [
{
"id": "flowpast-topbar-2643",
"name": "Flowpast Branding",
"type": "n8n-nodes-base.stickyNote",
"position": [
0,
20
],
"parameters": {
"color": 7,
"width": 1040,
"height": 80,
"content": "## Flowpast.com | Automation Workflow Library\n**\ud83d\udcd6 Full tutorial & setup guide:** flowpast.com"
},
"typeVersion": 1
},
{
"id": "f2fc02f9-a78a-4e87-be85-0032492a9f3f",
"name": "Current Time Stamp",
"type": "n8n-nodes-base.dateTime",
"position": [
730,
130
],
"parameters": {
"options": []
},
"typeVersion": 2
},
{
"id": "6f18ebbd-83db-4900-bc2e-0a9f23d6e8c8",
"name": "Incoming Request Trigger",
"type": "n8n-nodes-base.webhook",
"position": [
940,
140
],
"webhookId": "962f1468-c80f-4c0c-8555-a0acf648ede4",
"parameters": {
"path": "962f1468-c80f-4c0c-8555-a0acf648ede4",
"options": [],
"responseMode": "responseNode"
},
"typeVersion": 2
},
{
"id": "ba5ea83e-1b47-475b-863f-269ae293729a",
"name": "Auto Repair Parser B",
"type": "@n8n/n8n-nodes-langchain.outputParserAutofixing",
"position": [
700,
380
],
"parameters": {
"options": []
},
"typeVersion": 1
},
{
"id": "ca426b6d-5412-4c5b-a55c-009a47c59a81",
"name": "Auto Repair Parser A",
"type": "@n8n/n8n-nodes-langchain.outputParserAutofixing",
"position": [
240,
160
],
"parameters": {
"options": []
},
"typeVersion": 1
},
{
"id": "501d5390-5317-4973-a3e9-b0f502399c2b",
"name": "Schema Parser Alpha",
"type": "@n8n/n8n-nodes-langchain.outputParserStructured",
"position": [
100,
130
],
"parameters": {
"jsonSchemaExample": "{\n  \"reasoning_summary\": \"Detailed explanation of each analytical chain\u2019s purpose and insights, including key terms and considerations for query formulation.\",\n  \"final_search_query\": \"The single, best-fit search query derived from the meta-reasoning and multi-chain analysis, optimized to answer the research question.\"\n}"
},
"typeVersion": 1.2
},
{
"id": "a27e75c7-0307-4d71-9266-5a56b297a6e3",
"name": "Aggregate Search Results",
"type": "n8n-nodes-base.code",
"position": [
520,
320
],
"parameters": {
"jsCode": "// Initialize an empty string to store all title, url, and description pairs\nlet aggregatedOutputText = \"\";\n\n// Loop through all items passed to this Function node\nfor (let item of items) {\n  // Access the JSON data from \"Query 1\" node for the current item\n  const queryData = item.json;\n\n  // Ensure there is a \"web.results\" array to process\n  if (queryData.web?.results && Array.isArray(queryData.web.results)) {\n    // Loop through all results in the \"web.results\" array\n    for (let result of queryData.web.results) {\n      // Extract the title, url, and description for each result\n      const title = result.title || \"No Title\";\n      const url = result.url || \"No URL\";\n      const description = result.description || \"No Description\";\n\n      // Append the values to the aggregated string\n      aggregatedOutputText += `Title: ${title}\\nURL: ${url}\\nDescription: ${description}\\n\\n`;\n    }\n  } else {\n    // If no results array, handle gracefully\n    aggregatedOutputText += \"No results found for this item.\\n\\n\";\n  }\n}\n\n// Trim the final string to remove any trailing newline and whitespace\naggregatedOutputText = aggregatedOutputText.trim();\n\n// Return a single item containing the aggregated output as a string\nreturn [\n  {\n    json: {\n      aggregated_text: aggregatedOutputText\n    }\n  }\n];\n"
},
"typeVersion": 2
},
{
"id": "acbdbe94-b5a7-4ec9-9fc8-c3ab147f42fa",
"name": "Return Webhook Response",
"type": "n8n-nodes-base.respondToWebhook",
"position": [
940,
330
],
"parameters": {
"options": [],
"respondWith": "text",
"responseBody": "={\n    \"Highest_RANKEDURL_1\": {\n        \"title\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_1']['title'] }}\",\n        \"link\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_1']['link'] }}\",\n        \"description\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_1']['description'] }}\"\n    },\n    \"Highest_RANKEDURL_2\": {\n        \"title\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_2']['title'] }}\",\n        \"link\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_2']['link'] }}\",\n        \"description\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_2']['description'] }}\"\n    },\n    \"Highest_RANKEDURL_3\": {\n        \"title\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_3']['title'] }}\",\n        \"link\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_3']['link'] }}\",\n        \"description\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_3']['description'] }}\"\n    },\n    \"Highest_RANKEDURL_4\": {\n        \"title\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_4']['title'] }}\",\n        \"link\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_4']['link'] }}\",\n        \"description\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_4']['description'] }}\"\n    },\n    \"Highest_RANKEDURL_5\": {\n        \"title\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_5']['title'] }}\",\n        \"link\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_5']['link'] }}\",\n        \"description\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_5']['description'] }}\"\n    },\n    \"Highest_RANKEDURL_6\": {\n        \"title\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_6']['title'] }}\",\n        \"link\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_6']['link'] }}\",\n        \"description\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_6']['description'] }}\"\n    },\n    \"Highest_RANKEDURL_7\": {\n        \"title\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_7']['title'] }}\",\n        \"link\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_7']['link'] }}\",\n        \"description\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_7']['description'] }}\"\n    },\n    \"Highest_RANKEDURL_8\": {\n        \"title\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_8']['title'] }}\",\n        \"link\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_8']['link'] }}\",\n        \"description\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_8']['description'] }}\"\n    },\n    \"Highest_RANKEDURL_9\": {\n        \"title\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_9']['title'] }}\",\n        \"link\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_9']['link'] }}\",\n        \"description\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_9']['description'] }}\"\n    },\n    \"Highest_RANKEDURL_10\": {\n        \"title\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_10']['title'] }}\",\n        \"link\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_10']['link'] }}\",\n        \"description\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Highest_RANKEDURL_10']['description'] }}\"\n    },\n    \"Information_extracted\": \"{{ $item('0').$node['Semantic Result Reorder'].json['output']['Information_extracted'] }}\"\n}\n"
},
"typeVersion": 1.1
},
{
"id": "b8b6ae73-586a-406f-9641-57e2625f800c",
"name": "Semantic Result Reorder",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"onError": "continueRegularOutput",
"position": [
730,
300
],
"parameters": {
"text": "=\n**Objective:**\n\nFor the user's query, web search results are provided. Your tasks are:\n\n1. **Rank the links** based on how well they match the user's query.\n2. **Extract relevant information** from the descriptions provided. If no relevant information is found, return \"N/A\".\n\n---\n\n**Task:**\n\n**Step 1: Understand the User's Intent**\n\n- Determine what the user is truly and technically looking for.\n- The user's request query is: \"{{ $('Incoming Request Trigger').item.json.query['Research Question'] }}\"\n- The serach results below, however their  performance seem, have been based on this query \"{{ $item(\"0\").$node[\"Search Query Composer\"].json[\"output\"][\"final_search_query\"] }}\". If the result are not satisfactory or missing due to bad query making, you should note that as well for the neww query making.\n- To nesure being time aware , realize todays date is: \"{{ $item(\"0\").$node[\"Current Time Stamp\"].json[\"currentDate\"] }}\"\n\n- Follow a three-step chain of thought to comprehend the user's needs. Think out loud.\n\n---\n\n**Step 2: Rank the Links**\n\n- From the URLs and description snippets provided, **rank the top 10 websites** that are most likely to contain the required information.\n- Use the titles, descriptions, and sources to inform your ranking.\n\n**Links, Titles, and Descriptions:**\n\n{{ $json.aggregated_text }}\n\n---\n\nThis list completes the structure up to 20 results as you requested. Let me know if there\u2019s anything more you need!\n\n---\n\n**Step 3: Analyze and Create a Follow-up Query**\n\n- Recognize that for the user's request:\n\n  `\"{{ $('Incoming Request Trigger').item.json.query['Research Question'] }}\"`\n\n  The results provided are based on the assistant's generated search query:\n\n  `\"{{ $item(\"0\").$node[\"Search Query Composer\"].json[\"output\"][\"final_search_query\"] }}\"`\n\n- Analyze and revise any issues or new insights through multi-step thinking to create a follow-up query.\n\n**Indications and Priorities:**\n\n1. **No Results Received:** If no search items are shared, the search query may have been ineffective (e.g., too specific, incorrect parameters).\n2. **Insufficient or Unpromising Results:** If fewer than 20 but more than 5 results are provided, and none seem promising, the search query may need refinement.\n3. **Successful Results with Potential Follow-up:** If none of the above issues occurred and the search results provide answers or suggest a follow-up, create a new query. This could be a new topic, a deep dive, or a parallel factor that offers additional benefits.\n\n- Provide your chain of thought that connects the user's request to the actual information.\n\n- Deliver precise, detailed, and value-oriented information relevant to the user's query.\n\n**Step 4: Query making notes and examples**: \n\nThe queries must not be long tails , as they result in 0 websearch reutrns. We give you some examples of good web search queries:\nExamples:\n\nUser Question: \"What is the current state of the U.S. economy in 2024?\"\nEffective Search Query: \"U.S. Economy Analysis Report 2024\"\n\nUser Question: \"What are the recent advancements in artificial intelligence?\"\nEffective Search Query: \"2024 Artificial Intelligence Developments\"\n\nUser Question: \"How is climate change affecting agriculture globally?\"\nEffective Search Query: \"Global Impact of Climate Change on Agriculture 2024\"\n\nUser Question: \"What are the latest trends in cybersecurity threats?\"\nEffective Search Query: \"Cybersecurity Threats and Trends 2024\"\n\nUser Question: \"What is the outlook for renewable energy investments?\"\nEffective Search Query: \"Renewable Energy Investment Outlook 2024\"\n\n**Step 5: Query making*: \nor query making remember as we said:\n   - **Today's Date:** \"{{ $item(\"0\").$node[\"Current Time Stamp\"].json[\"currentDate\"] }}\"\n **Search Inquiry:** \n   - **Search Topic to create the query upon it:**{{ $item(\"0\").$node[\"Incoming Request Trigger\"].json[\"query\"][\"Research Question\"] }}\"\"\n\n\n---\n\n**Step 6: Output Format**\n\nEnsure the response is in the following JSON format:\n\n{\n    \"chain_of_thought\": \"Insert your step-by-step reasoning here.\",\n    \"Highest_RANKEDURL_1\": {\n        \"title\": \"Insert the First Ranked URL's Title here.\",\n        \"link\": \"Insert the First Ranked URL here.\",\n        \"description\": \"Insert the First Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_2\": {\n        \"title\": \"Insert the Second Ranked URL's Title here.\",\n        \"link\": \"Insert the Second Ranked URL here.\",\n        \"description\": \"Insert the Second Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_3\": {\n        \"title\": \"Insert the Third Ranked URL's Title here.\",\n        \"link\": \"Insert the Third Ranked URL here.\",\n        \"description\": \"Insert the Third Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_4\": {\n        \"title\": \"Insert the Fourth Ranked URL's Title here.\",\n        \"link\": \"Insert the Fourth Ranked URL here.\",\n        \"description\": \"Insert the Fourth Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_5\": {\n        \"title\": \"Insert the Fifth Ranked URL's Title here.\",\n        \"link\": \"Insert the Fifth Ranked URL here.\",\n        \"description\": \"Insert the Fifth Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_6\": {\n        \"title\": \"Insert the Sixth Ranked URL's Title here.\",\n        \"link\": \"Insert the Sixth Ranked URL here.\",\n        \"description\": \"Insert the Sixth Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_7\": {\n        \"title\": \"Insert the Seventh Ranked URL's Title here.\",\n        \"link\": \"Insert the Seventh Ranked URL here.\",\n        \"description\": \"Insert the Seventh Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_8\": {\n        \"title\": \"Insert the Eighth Ranked URL's Title here.\",\n        \"link\": \"Insert the Eighth Ranked URL here.\",\n        \"description\": \"Insert the Eighth Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_9\": {\n        \"title\": \"Insert the Ninth Ranked URL's Title here.\",\n        \"link\": \"Insert the Ninth Ranked URL here.\",\n        \"description\": \"Insert the Ninth Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_10\": {\n        \"title\": \"Insert the Tenth Ranked URL's Title here.\",\n        \"link\": \"Insert the Tenth Ranked URL here.\",\n        \"description\": \"Insert the Tenth Ranked URL's Description here.\"\n    },\n    \"Information_extracted\": \"Insert all extracted information relevant to the user's query or 'N/A' if none.\"\n}\n",
"messages": {
"messageValues": [
{
"message": "=\nYou are an expert information retrieval and critical evaluation assistant designed to process, rank, and extract high-relevance content from web search results for complex user queries. You must provide value-oriented insights while refining searches based on relevance and context sensitivity. \n\n**Your Process and Priorities:**\n\n#### 1. **Determine the User's Technical Intent**\n   - Interpret the user's core question provided as `{{ $item(\"0\").$node[\"Incoming Request Trigger\"].json[\"query\"][\"Research Question\"] }}`, discerning underlying objectives and specialized needs.\n   - Recognize that the search results may have been generated from a **secondary query**: `{{ $item(\"0\").$node[\"Search Query Composer\"].json[\"output\"][\"final_search_query\"] }}`. \n   - Judge the adequacy of this generated query. If it does not meet the user\u2019s objectives, highlight the need for query refinement and prepare to adapt the approach.\n   - Stay mindful of the date context, using `{{ $item(\"0\").$node[\"Current Time Stamp\"].json[\"currentDate\"] }}` to assess the freshness of content or time-sensitive relevance.\n\n#### 2. **Rank Results Based on Analytical Relevance**\n   - From the search results provided, **rank the top 3 URLs** that most closely align with the user\u2019s intent and technical needs.\n     - Use multi-dimensional analysis to assess how each link\u2019s title, description, and source match the user\u2019s objective.\n     - Prioritize results based on credibility, relevance, and their potential to add depth to the user\u2019s inquiry.\n   - Your goal is to select the highest-value links, disregarding results that offer superficial, off-topic, or outdated information.\n\n#### 3. **Extract Key Information**\n   - For each of the top 3 ranked results, extract insights and details from the description snippets that directly address the user\u2019s query.\n   - If no pertinent information is available in a description, record `\"N/A\"` to indicate its lack of relevance.\n\n#### 4. **Evaluate for Potential Query Improvement**\n   - Evaluate the relevance and coverage of search results:\n     - If fewer than 5 relevant results are present, consider that the initial query may be too narrow, specific, or otherwise misaligned.\n     - Generate a **refined query** that is adjusted to better match the user\u2019s likely needs and produce higher-quality results.\n   - Use advanced language modifications, new keyword suggestions, or rephrasing to formulate a search query that enhances alignment with the user\u2019s goals.\n"
}
]
},
"promptType": "define",
"hasOutputParser": true
},
"retryOnFail": true,
"typeVersion": 1.4
},
{
"id": "a1ca671d-0b0c-4717-9def-93fdb965de8d",
"name": "Primary Web Search Call",
"type": "n8n-nodes-base.httpRequest",
"position": [
300,
130
],
"parameters": {
"url": "https://api.search.brave.com/res/v1/web/search",
"options": [],
"sendQuery": true,
"sendHeaders": true,
"queryParameters": {
"parameters": [
{
"name": "q",
"value": "={{ $json.output.final_search_query }}"
}
]
},
"headerParameters": {
"parameters": [
{
"name": "Accept",
"value": "application/json"
},
{
"name": "Accept-Encoding",
"value": "gzip"
},
{
"name": "X-Subscription-Token",
"value": "[CONFIGURE_YOUR_API_KEY]"
}
]
}
},
"typeVersion": 4.2
},
{
"id": "d3cc4e7c-3ead-4d38-9b51-a11cd9d7faeb",
"name": "Utility: Webhook Test Call",
"type": "n8n-nodes-base.httpRequest",
"position": [
0,
520
],
"parameters": {
"url": "https://example.com/webhook/[YOUR_ID]",
"options": [],
"sendQuery": true,
"queryParameters": {
"parameters": [
{
"name": "Research Question",
"value": "what is the latest news in global world in politics and economy?"
}
]
}
},
"typeVersion": 4.2
},
{
"id": "01d73f91-1dd6-4b80-951c-9f944ea9d992",
"name": "Search Query Composer",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
520,
150
],
"parameters": {
"text": "=1. **Task:** `\"Your task is to develop a web search query that most effectively answers the research question given. Use meta-reasoning and multi-chain analysis to ensure a comprehensive approach.\"`\n\n2. **Structured Guidance for Chains of Thought:**  \n    a. **Chain 1:** Break down the research question, identifying keywords and relevant terms.  \n    b. **Chain 2:** Explore the context and potential sources, determining the types of results that would be most relevant.  \n    c. **Chain 3:** Refine the query for specificity and completeness, considering how to capture nuances of the question.\n\n3. **Final Query Generation:** Based on the insights from the three chains, generate a single, refined search query.\n\n\n4. Note, the queries must not be long tails , as they result in 0 websearch reutrns. We give you some examples of good web search queries:\nExamples:\n\nUser Question: \"What is the current state of the U.S. economy in 2024?\"\n\nEffective Search Query: \"U.S. Economy Analysis Report 2024\"\nUser Question: \"What are the recent advancements in artificial intelligence?\"\n\nEffective Search Query: \"2024 Artificial Intelligence Developments\"\nUser Question: \"How is climate change affecting agriculture globally?\"\n\nEffective Search Query: \"Global Impact of Climate Change on Agriculture 2024\"\nUser Question: \"What are the latest trends in cybersecurity threats?\"\n\nEffective Search Query: \"Cybersecurity Threats and Trends 2024\"\nUser Question: \"What is the outlook for renewable energy investments?\"\n\nEffective Search Query: \"Renewable Energy Investment Outlook 2024\"\n\n5. Data Input:\n   - **Today's Date:** \"{{ $item(\"0\").$node[\"Current Time Stamp\"].json[\"currentDate\"] }}\"\n **Search Inquiry:** \n   - **Search Topic to create the query upon it:**{{ $item(\"0\").$node[\"Incoming Request Trigger\"].json[\"query\"][\"Research Question\"] }}\"\"\n\n6. Now develop the best fit web search query given the user request above under number 5\n---\n\n**Output Requirements:**  \nThe Assistant\u2019s output should be in JSON format, structured as follows:\n\n{\n  \"reasoning_summary\": \"Detailed explanation of each analytical chain\u2019s purpose and insights, including key terms and considerations for query formulation.\",\n  \"final_search_query\": \"The single, best-fit search query derived from the meta-reasoning and multi-chain analysis, optimized to answer the research question.\"\n}\n```\n\n---\n",
"messages": {
"messageValues": [
{
"message": "You are an advanced data and research retrieval through smart search queires via Bing and Brave websearch APIs. "
},
{
"type": "HumanMessagePromptTemplate",
"message": "1. **Task:** `\"Your task is to develop a web search query that most effectively answers the research question given. Use meta-reasoning and multi-chain analysis to ensure a comprehensive approach.\"`\n\n2. **Structured Guidance for Chains of Thought:**  \n    a. **Chain 1:** Break down the research question, identifying keywords and relevant terms.  \n    b. **Chain 2:** Explore the context and potential sources, determining the types of results that would be most relevant.  \n    c. **Chain 3:** Refine the query for specificity and completeness, considering how to capture nuances of the question.\n\n3. **Final Query Generation:** Based on the insights from the three chains, generate a single, refined search query.\n\n4. Data Input:\n   - **Today's Date:** \"2024-11-12T10:21:33.764-05:00\"\n2. **Search Inquiry:** \n   - **Search Topic to create the query upon it:** \"What is the latest stock Market Analysis in mid-term?\"\n\n\n---\n\n**Output Requirements:**  \nThe Assistant\u2019s output should be in JSON format, structured as follows:\n\n{\n  \"reasoning_summary\": \"Detailed explanation of each analytical chain\u2019s purpose and insights, including key terms and considerations for query formulation.\",\n  \"final_search_query\": \"The single, best-fit search query derived from the meta-reasoning and multi-chain analysis, optimized to answer the research question.\"\n}\n```\n\n---\n"
},
{
"type": "AIMessagePromptTemplate",
"message": "Assessing query functionality\n\nI\u2019m evaluating the viability of the assistant's proposed search query. The user questions if it would yield meaningful results on different search engines.\n\nClarifying the query\n\nOK, let me see. The task involves clarifying the user's needs, ensuring the response aligns with OpenAI's policies, and possibly re-evaluating and correcting any errors in the previous response.\n\nRefining search specifics\n\nI'm working through refining the search query to focus on specific terms like 'S&P 500', integrating dates such as 'November 2024', and emphasizing phrases for precision.\n\nSelf-reflecting on search queries\n\nToday, I\u2019m thinking about how to validate search queries for accuracy and realism, ensuring they are practical and likely to yield productive results.\n\nAcknowledging insights\n\nI\u2019m recognizing and correcting a search query to ensure it\u2019s optimal and effectively aligned with advanced user needs, leading to a more precise output.\n\nCrafting a search query\n\nI\u2019m crafting a search query to pinpoint the most pertinent recent mid-term stock market analyses. Starting with relevant terms and then exploring potential sources for context and relevance.\n\nPinpointing sources and specifics\n\nI'm gathering info from financial news, investment reports, and economic research institutions. I\u2019m refining search terms like \"mid-term stock market analysis\" and \"market forecast\" to keep results current and nuanced.\n\nCrafting the query\n\nI\u2019m finalizing the query by merging terms like 'latest' and 'stock market analysis' with 'mid-term', and considering synonyms such as 'market forecast' to refine the search for 2024.\n\nTweaking the search\n\nI adjusted the search terms to 'latest stock market mid-term outlook 2024,' removing excessive quotes for better clarity and relevance.\n\n\n{\n  \"reasoning_summary\": \"Chain 1: Broke down the research question to identify key terms such as 'latest', 'stock market analysis', and 'mid-term'. Considered synonyms like 'mid-term outlook', 'market forecast', and included the current year '2024' for recency.\\n\\nChain 2: Explored the context where the user seeks recent analyses of the stock market with a mid-term perspective. Relevant sources would be financial news outlets, investment banks, and economic research firms. The most useful results would be up-to-date articles or reports offering mid-term forecasts.\\n\\nChain 3: Refined the query by removing unnecessary quotation marks and simplifying it to reflect how an advanced user might search. Included terms like 'latest', 'mid-term outlook', and '2024' to enhance specificity without limiting the search results unnecessarily.\",\n  \"final_search_query\": \"latest stock market mid-term outlook 2024\"\n}"
}
]
},
"promptType": "define",
"hasOutputParser": true
},
"typeVersion": 1.4
},
{
"id": "696b4f97-ad29-406b-9157-44ad9d05c9cd",
"name": "Utility: Claude Chat Model",
"type": "@n8n/n8n-nodes-langchain.lmChatAnthropic",
"position": [
360,
520
],
"parameters": {
"model": "claude-3-5-haiku-20241022",
"options": {
"topP": 0.8,
"temperature": 0.4,
"maxTokensToSample": 4096
}
},
"credentials": {
"anthropicApi": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 1.2
},
{
"id": "82f25610-8b70-4aee-ad90-2616d3389f15",
"name": "Utility: OpenAI Dialogue Model",
"type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
"position": [
700,
520
],
"parameters": {
"options": {
"topP": 0.7,
"maxTokens": 4096,
"maxRetries": 1,
"temperature": 0.5
}
},
"credentials": {
"openAiApi": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 1
},
{
"id": "f230bdf0-4a22-4abf-96cd-47f309f0c514",
"name": "Schema Parser Beta",
"type": "@n8n/n8n-nodes-langchain.outputParserStructured",
"position": [
560,
420
],
"parameters": {
"jsonSchemaExample": "{\n    \"chain_of_thought\": \"Insert your step-by-step reasoning here.\",\n    \"Highest_RANKEDURL_1\": {\n        \"title\": \"Insert the First Ranked URL's Title here.\",\n        \"link\": \"Insert the First Ranked URL here.\",\n        \"description\": \"Insert the First Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_2\": {\n        \"title\": \"Insert the Second Ranked URL's Title here.\",\n        \"link\": \"Insert the Second Ranked URL here.\",\n        \"description\": \"Insert the Second Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_3\": {\n        \"title\": \"Insert the Third Ranked URL's Title here.\",\n        \"link\": \"Insert the Third Ranked URL here.\",\n        \"description\": \"Insert the Third Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_4\": {\n        \"title\": \"Insert the Fourth Ranked URL's Title here.\",\n        \"link\": \"Insert the Fourth Ranked URL here.\",\n        \"description\": \"Insert the Fourth Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_5\": {\n        \"title\": \"Insert the Fifth Ranked URL's Title here.\",\n        \"link\": \"Insert the Fifth Ranked URL here.\",\n        \"description\": \"Insert the Fifth Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_6\": {\n        \"title\": \"Insert the Sixth Ranked URL's Title here.\",\n        \"link\": \"Insert the Sixth Ranked URL here.\",\n        \"description\": \"Insert the Sixth Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_7\": {\n        \"title\": \"Insert the Seventh Ranked URL's Title here.\",\n        \"link\": \"Insert the Seventh Ranked URL here.\",\n        \"description\": \"Insert the Seventh Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_8\": {\n        \"title\": \"Insert the Eighth Ranked URL's Title here.\",\n        \"link\": \"Insert the Eighth Ranked URL here.\",\n        \"description\": \"Insert the Eighth Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_9\": {\n        \"title\": \"Insert the Ninth Ranked URL's Title here.\",\n        \"link\": \"Insert the Ninth Ranked URL here.\",\n        \"description\": \"Insert the Ninth Ranked URL's Description here.\"\n    },\n    \"Highest_RANKEDURL_10\": {\n        \"title\": \"Insert the Tenth Ranked URL's Title here.\",\n        \"link\": \"Insert the Tenth Ranked URL here.\",\n        \"description\": \"Insert the Tenth Ranked URL's Description here.\"\n    },\n    \"Information_extracted\": \"Insert all extracted information relevant to the user's query or 'N/A' if none.\"\n}\n"
},
"typeVersion": 1.2
},
{
"id": "3421ebe5-6a86-435e-b9c6-e3dcf6dd7833",
"name": "Gemini Parsing Model",
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"position": [
380,
240
],
"parameters": {
"options": {
"topP": 0.6,
"temperature": 0.4,
"maxOutputTokens": 4096
},
"modelName": "models/gemini-1.5-flash-002"
},
"credentials": {
"googlePalmApi": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 1
},
{
"id": "fd19e9f6-d8c7-45df-93d9-ecf7956b461f",
"name": "Gemini Agent Model",
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"position": [
620,
120
],
"parameters": {
"options": {
"topP": 0.6,
"temperature": 0.4,
"safetySettings": {
"values": [
{
"category": "HARM_CATEGORY_HARASSMENT",
"threshold": "BLOCK_NONE"
},
{
"category": "HARM_CATEGORY_HATE_SPEECH",
"threshold": "BLOCK_NONE"
},
{
"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",
"threshold": "BLOCK_NONE"
},
{
"category": "HARM_CATEGORY_DANGEROUS_CONTENT",
"threshold": "BLOCK_NONE"
}
]
},
"maxOutputTokens": 4086
},
"modelName": "models/gemini-1.5-flash-002"
},
"credentials": {
"googlePalmApi": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 1
}
],
"active": false,
"pinData": [],
"settings": {
"executionOrder": "v1"
},
"versionId": null,
"connections": {
"Primary Web Search Call": {
"main": [
[
{
"node": "Aggregate Search Results",
"type": "main",
"index": 0
}
]
]
},
"Incoming Request Trigger": {
"main": [
[
{
"node": "Current Time Stamp",
"type": "main",
"index": 0
}
]
]
},
"Gemini Agent Model": {
"ai_languageModel": [
[
{
"node": "Semantic Result Reorder",
"type": "ai_languageModel",
"index": 0
},
{
"node": "Search Query Composer",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Current Time Stamp": {
"main": [
[
{
"node": "Search Query Composer",
"type": "main",
"index": 0
}
]
]
},
"Gemini Parsing Model": {
"ai_languageModel": [
[
{
"node": "Auto Repair Parser B",
"type": "ai_languageModel",
"index": 0
},
{
"node": "Auto Repair Parser A",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Aggregate Search Results": {
"main": [
[
{
"node": "Semantic Result Reorder",
"type": "main",
"index": 0
}
]
]
},
"Utility: OpenAI Dialogue Model": {
"ai_languageModel": [
[]
]
},
"Utility: Claude Chat Model": {
"ai_languageModel": [
[]
]
},
"Auto Repair Parser A": {
"ai_outputParser": [
[
{
"node": "Search Query Composer",
"type": "ai_outputParser",
"index": 0
}
]
]
},
"Schema Parser Alpha": {
"ai_outputParser": [
[
{
"node": "Auto Repair Parser A",
"type": "ai_outputParser",
"index": 0
}
]
]
},
"Schema Parser Beta": {
"ai_outputParser": [
[
{
"node": "Auto Repair Parser B",
"type": "ai_outputParser",
"index": 0
}
]
]
},
"Auto Repair Parser B": {
"ai_outputParser": [
[
{
"node": "Semantic Result Reorder",
"type": "ai_outputParser",
"index": 0
}
]
]
},
"Search Query Composer": {
"main": [
[
{
"node": "Primary Web Search Call",
"type": "main",
"index": 0
}
]
]
},
"Semantic Result Reorder": {
"main": [
[
{
"node": "Return Webhook Response",
"type": "main",
"index": 0
}
]
]
}
}
}