CV Resume PDF Parsing with Multimodal Vision AI.json

{
"meta": {
"instanceId": "408f9fb9940c3cb18ffdef0e0150fe342d6e655c3a9fac21f0f644e8bedabcd9"
},
"nodes": [
{
"id": "38da57b7-2161-415d-8473-783ccdc7b975",
"name": "When clicking ‘Test workflow’",
"type": "n8n-nodes-base.manualTrigger",
"position": [
-260,
840
],
"parameters": {},
"typeVersion": 1
},
{
"id": "2cd46d91-105d-4b5e-be43-3343a9da815d",
"name": "Sticky Note",
"type": "n8n-nodes-base.stickyNote",
"position": [
-780,
540
],
"parameters": {
"width": 365.05232558139534,
"height": 401.24529475392126,
"content": "## Try me out!\n\n### This workflow converts a Candidate Resume PDF to an image which is then \"read\" by a Vision Language Model (VLM). The VLM assesses if the candidate's CV is a fit for the desired role.\n\nThis approach can be employed to combat \"hidden prompts\" planted in resumes to bypass and/or manipulate automated ATS systems using AI.\n\n\n### Need Help?\nJoin the [Discord](https://discord.com/invite/XPKeKXeB7d) or ask in the [Forum](https://community.n8n.io/)!\n"
},
"typeVersion": 1
},
{
"id": "40bab53a-fcbc-4acc-8d59-c20b3e1b2697",
"name": "Structured Output Parser",
"type": "@n8n/n8n-nodes-langchain.outputParserStructured",
"position": [
1200,
980
],
"parameters": {
"jsonSchemaExample": "{\n\t\"is_qualified\": true,\n\t\"reason\": \"\"\n}"
},
"typeVersion": 1.2
},
{
"id": "d75fb7ab-cfbc-419d-b803-deb9e99114ba",
"name": "Should Proceed To Stage 2?",
"type": "n8n-nodes-base.if",
"position": [
1360,
820
],
"parameters": {
"options": {},
"conditions": {
"options": {
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "4dd69ba3-bf07-43b3-86b7-d94b07e9eea6",
"operator": {
"type": "boolean",
"operation": "true",
"singleValue": true
},
"leftValue": "={{ $json.output.is_qualified }}",
"rightValue": ""
}
]
}
},
"typeVersion": 2
},
{
"id": "a0f56270-67c2-4fab-b521-aa6f06b0b0fd",
"name": "Sticky Note1",
"type": "n8n-nodes-base.stickyNote",
"position": [
-380,
540
],
"parameters": {
"color": 7,
"width": 543.5706868577606,
"height": 563.6162790697684,
"content": "## 1. Download Candidate Resume\n[Read more about using Google Drive](https://docs.n8n.io/integrations/builtin/app-nodes/n8n-nodes-base.googledrive)\n\nFor this demonstration, we'll pull the candidate's resume PDF from Google Drive but you can just as easily recieve this resume from email or your ATS.\n\nIt should be noted that our PDF is a special test case which has been deliberately injected with an AI bypass; the bypass is a hidden prompt which aims to override AI instructions and auto-qualify the candidate... sneaky!\n\nDownload a copy of this resume here: https://drive.google.com/file/d/1MORAdeev6cMcTJBV2EYALAwll8gCDRav/view?usp=sharing"
},
"typeVersion": 1
},
{
"id": "d21fe4dd-0879-4e5a-a70d-10f09b25eee2",
"name": "Download Resume",
"type": "n8n-nodes-base.googleDrive",
"position": [
-80,
840
],
"parameters": {
"fileId": {
"__rl": true,
"mode": "id",
"value": "1MORAdeev6cMcTJBV2EYALAwll8gCDRav"
},
"options": {},
"operation": "download"
},
"credentials": {
"googleDriveOAuth2Api": {
"id": "yOwz41gMQclOadgu",
"name": "Google Drive account"
}
},
"typeVersion": 3
},
{
"id": "ea904365-d9d2-4f15-b7c3-7abfeb4c8c50",
"name": "Sticky Note2",
"type": "n8n-nodes-base.stickyNote",
"position": [
200,
540
],
"parameters": {
"color": 7,
"width": 605.0267171444024,
"height": 595.3148729042731,
"content": "## 2. Convert PDF to Image(s)\n[Read more about using Stirling PDF](https://github.com/Stirling-Tools/Stirling-PDF)\n\nAI vision models can only accept images (and sometimes videos!) as non-text inputs but not PDFs at time of writing. We'll have to convert our PDF to an image in order to use it.\n\nHere, we'll use a tool called **Stirling PDF** which can provide this functionality and can be accessed via a HTTP API. Feel free to use an alternative solution if available, otherwise follow the instructions on the Stirling PDF website to set up your own instance.\n\nAdditionally, we'll reduce the resolution of our converted image to speed up the processing done by the LLM. I find that about 75% of an A4 (30x40cm) is a good balance."
},
"typeVersion": 1
},
{
"id": "cd00a47f-1ab9-46bf-8ea1-46ac899095e7",
"name": "Sticky Note3",
"type": "n8n-nodes-base.stickyNote",
"position": [
840,
540
],
"parameters": {
"color": 7,
"width": 747.8139534883712,
"height": 603.1395348837208,
"content": "## 3. Parse Resume with Multimodal LLM\n[Read more about using Basic LLM Chain](https://docs.n8n.io/integrations/builtin/cluster-nodes/root-nodes/n8n-nodes-langchain.chainllm/)\n\nMultimodal LLMs are LLMs which can accept binary inputs such as images, audio and/or video files. Most newer LLMs are by default multimodal and we'll use Google's Gemini here as an example. By processing each candidate's resume as an image, we avoid scenarios where text extraction fails due to layout issues or by picking up \"hidden\" or malicious prompts planted to subvert AI automated processing.\n\nThis vision model ensures the resume is read and understood as a human would. The hidden bypass is therefore rendered mute since the AI also cannot \"see\" the special prompt embedded in the document."
},
"typeVersion": 1
},
{
"id": "d60214c6-c67e-4433-9121-4d54f782b19d",
"name": "PDF-to-Image API",
"type": "n8n-nodes-base.httpRequest",
"position": [
340,
880
],
"parameters": {
"url": "https://stirlingpdf.io/api/v1/convert/pdf/img",
"method": "POST",
"options": {},
"sendBody": true,
"contentType": "multipart-form-data",
"bodyParameters": {
"parameters": [
{
"name": "fileInput",
"parameterType": "formBinaryData",
"inputDataFieldName": "data"
},
{
"name": "imageFormat",
"value": "jpg"
},
{
"name": "singleOrMultiple",
"value": "single"
},
{
"name": "dpi",
"value": "300"
}
]
}
},
"typeVersion": 4.2
},
{
"id": "847de537-ad8f-47f5-a1c1-d207c3fc15ef",
"name": "Resize Converted Image",
"type": "n8n-nodes-base.editImage",
"position": [
530,
880
],
"parameters": {
"width": 75,
"height": 75,
"options": {},
"operation": "resize",
"resizeOption": "percent"
},
"typeVersion": 1
},
{
"id": "5fb6ac7e-b910-4dce-bba7-19b638fd817a",
"name": "Google Gemini Chat Model",
"type": "@n8n/n8n-nodes-langchain.lmChatGoogleGemini",
"position": [
1000,
980
],
"parameters": {
"options": {},
"modelName": "models/gemini-1.5-pro-latest"
},
"credentials": {
"googlePalmApi": {
"id": "dSxo6ns5wn658r8N",
"name": "Google Gemini(PaLM) Api account"
}
},
"typeVersion": 1
},
{
"id": "2580b583-544a-47ee-b248-9cca528c9866",
"name": "Candidate Resume Analyser",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
1000,
820
],
"parameters": {
"text": "=Evaluate the candidate's resume.",
"messages": {
"messageValues": [
{
"message": "=Assess the given Candiate Resume for the role of Plumber.\nDetermine if the candidate's skills match the role and if they qualify for an in-person interview."
},
{
"type": "HumanMessagePromptTemplate",
"messageType": "imageBinary"
}
]
},
"promptType": "define",
"hasOutputParser": true
},
"typeVersion": 1.4
},
{
"id": "694669c2-9cf5-43ec-8846-c0ecbc5a77ee",
"name": "Sticky Note4",
"type": "n8n-nodes-base.stickyNote",
"position": [
280,
840
],
"parameters": {
"width": 225.51725256895617,
"height": 418.95152406706313,
"content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n### Data Privacy Warning!\nFor demo purposes, we're using the public online version of Stirling PDF. It is recommended to setup your own private instance of Stirling PDF before using this workflow in production."
},
"typeVersion": 1
}
],
"pinData": {},
"connections": {
"Download Resume": {
"main": [
[
{
"node": "PDF-to-Image API",
"type": "main",
"index": 0
}
]
]
},
"PDF-to-Image API": {
"main": [
[
{
"node": "Resize Converted Image",
"type": "main",
"index": 0
}
]
]
},
"Resize Converted Image": {
"main": [
[
{
"node": "Candidate Resume Analyser",
"type": "main",
"index": 0
}
]
]
},
"Google Gemini Chat Model": {
"ai_languageModel": [
[
{
"node": "Candidate Resume Analyser",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Structured Output Parser": {
"ai_outputParser": [
[
{
"node": "Candidate Resume Analyser",
"type": "ai_outputParser",
"index": 0
}
]
]
},
"Candidate Resume Analyser": {
"main": [
[
{
"node": "Should Proceed To Stage 2?",
"type": "main",
"index": 0
}
]
]
},
"When clicking ‘Test workflow’": {
"main": [
[
{
"node": "Download Resume",
"type": "main",
"index": 0
}
]
]
}
}
}