January 22, 2026

Gmail + Google Sheets: resumes scored and logged

Lisa Granqvist Partner Workflow Automation Expert

Get a free AI assessment → ⬇️ Use template

Your inbox shouldn’t be your applicant tracking system. But when resumes arrive as PDFs in Gmail, it’s easy to lose track, forget who applied for what role, or end up “saving” files in five different places.

This Gmail resume scoring automation hits recruiters first, honestly. HR generalists and founders hiring between meetings feel it too. The outcome is simple: every resume gets stored, scored, and logged so you can review top fits fast without manual copy-paste.

Below, you’ll see exactly how the workflow moves from a new email to a clean Google Sheet row, plus what you need to run it and the common pitfalls to avoid.

How This Automation Works

The full n8n workflow, from trigger to final output:

n8n Workflow Template: Gmail + Google Sheets: resumes scored and logged

Click to explore

flowchart LR

    subgraph sg0["Inbox Attachment Watch Flow"]
        direction LR
        n0@{ icon: "mdi:play-circle", form: "rounded", label: "Inbox Attachment Watch", 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/airtable.svg' width='40' height='40' /></div><br/>Retrieve Job Listings"]
        n2["<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/>Match Job From Subject"]
        n3@{ icon: "mdi:swap-horizontal", form: "rounded", label: "Match Found Check", pos: "b", h: 48 }
        n4["<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/>Compose AI Prompt"]
        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/merge.svg' width='40' height='40' /></div><br/>Initial Stream Merge"]
        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/code.svg' width='40' height='40' /></div><br/>Prepare Resume Files"]
        n7["<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/>Upload Resume File"]
        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/merge.svg' width='40' height='40' /></div><br/>Unify Prompt And File"]
        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/code.svg' width='40' height='40' /></div><br/>Assemble Gemini Payload"]
        n10["<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/>Request AI Analysis"]
        n11["<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/merge.svg' width='40' height='40' /></div><br/>Aggregate Result Data"]
        n12@{ icon: "mdi:cog", form: "rounded", label: "Archive To Drive", pos: "b", h: 48 }
        n13["<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/>Parse AI Output"]
        n14@{ icon: "mdi:database", form: "rounded", label: "Append To Sheet", pos: "b", h: 48 }
        n3 --> n5
        n3 --> n4
        n8 --> n9
        n5 --> n6
        n1 --> n2
        n7 --> n8
        n11 --> n13
        n2 --> n3
        n12 --> n11
        n0 --> n1
        n0 --> n5
        n13 --> n14
        n4 --> n8
        n4 --> n11
        n9 --> n10
        n6 --> n7
        n6 --> n12
        n6 --> n11
        n10 --> n11
    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 n0 trigger
    class n3 decision
    class n1,n14 database
    class n7,n10 api
    class n2,n4,n6,n9,n13 code
    classDef customIcon fill:none,stroke:none
    class n1,n2,n4,n5,n6,n7,n8,n9,n10,n11,n13 customIcon

The Problem: Resume screening turns into inbox chaos

When applications come in by email, the process looks “fine” until you’re buried. One candidate emails twice with updated PDFs. Another forgets to include the job title. You download attachments to your desktop, then re-upload them later, then try to remember which spreadsheet row matches which file. Meanwhile, hiring managers ask, “Any good ones yet?” and you’re still sorting subject lines. The real cost is not just time. It’s missed follow-ups, inconsistent evaluation, and decisions made from incomplete notes.

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

Someone has to read each email subject and guess which open role it belongs to.
Resume files get downloaded and re-saved, which creates duplicates and broken links.
Scoring is inconsistent because it depends on who had time to review that day.
Google Sheets updates lag behind reality, so the “pipeline view” can’t be trusted.

The Solution: Automatically score Gmail resumes and log them to Sheets

This workflow watches your Gmail inbox for new emails that include resume attachments (PDFs). As soon as one arrives, it pulls the job title from the email subject and checks it against your open positions in Airtable, so the resume gets evaluated in the right context. The resume file is then stored in a designated Google Drive folder (so you have one clean source of truth), and the workflow sends both the file and a structured prompt to Gemini for analysis. Gemini returns a score, extracts key candidate details like name and phone, and adds a one-sentence explanation you can scan quickly. Finally, n8n appends everything into Google Sheets, including a download link to the stored resume.

The flow starts with Gmail, then Airtable provides job context, then Gemini produces a consistent evaluation. Google Drive stores the file, and Google Sheets becomes the “single list” your team can trust.

What You Get: Automation vs. Results

What This Workflow Automates

Results You’ll Get

Monitors Gmail for new resume emails with PDF attachments.
Matches the email subject to open job listings stored in Airtable.
Stores each resume in Google Drive and keeps the link attached.
Sends the prompt and file to Gemini, then parses the AI output into fields.

Most teams get about 5 hours back per week once volume picks up.
Every candidate is scored the same way, so comparisons are fair.
Your spreadsheet becomes a live pipeline view instead of a delayed report.
Fewer “where’s that resume?” pings because the file link is always there.
You can shortlist faster because the explanation line makes scanning easy.

Example: What This Looks Like

Say you get 20 resumes in a week across 3 open roles. Manually, you might spend about 10 minutes per resume downloading the PDF, renaming it, filing it in Drive, checking the role, and adding notes to a sheet, which is roughly 3 hours weekly. With this workflow, the “work” is basically opening the sheet later: Gmail triggers instantly, Drive storage and Gemini scoring run in the background, and the row appears automatically. You get your 3 hours back, plus a cleaner shortlist.

What You’ll Need

n8n instance (try n8n Cloud free)
Self-hosting option if you prefer (Hostinger works well)
Gmail to receive applications and trigger runs
Google Drive to store resumes in one folder
Google Sheets to log candidates and AI scores
Airtable for job listings and descriptions
Gemini API key (get it from Google AI Studio)

Skill level: Intermediate. You’ll connect accounts, map a few fields, and paste in your job descriptions.

Don’t want to set this up yourself? Talk to an automation expert (free 15-minute consultation).

How It Works

New resume email hits your inbox. The workflow triggers from Gmail when a message arrives with an attachment, so nothing relies on someone remembering to check a folder.

Job context is pulled from Airtable. n8n fetches your open job listings and uses the email subject to match the candidate to the right role. If there’s no match, the workflow can stop early instead of scoring against the wrong job.

The resume is stored and prepared for AI. The PDF is uploaded and archived in Google Drive, and n8n composes a clear prompt that includes the job description and the evaluation criteria you care about.

Gemini scores and extracts the details. The workflow sends the resume file plus the prompt via HTTP Request, then parses Gemini’s response into structured fields like name, phone, score, and a short explanation.

Everything is appended to Google Sheets. A new row is created with candidate info, the AI score, feedback, and a Drive download link so review and sharing are frictionless.

You can easily modify the matching logic to use job codes instead of subject titles based on your needs. See the full implementation guide below for customization options.

Step-by-Step Implementation Guide

Step 1: Configure the Gmail Trigger

Set up the workflow to watch for incoming emails with attachments so resumes are captured immediately.

Add the Inbox Attachment Watch node as your trigger.
Set Filters → q to has:attachment.
Enable Options → downloadAttachments so the files are available downstream.
Set Poll Times → item → mode to everyMinute for near-real-time ingestion.
Credential Required: Connect your gmailOAuth2 credentials.

Step 2: Connect Airtable and Match Jobs

Pull job listings and match the email subject to the most relevant role.

Add Retrieve Job Listings and set Operation to search.
Choose the Airtable Base and Table from your environment (e.g., “ResumeRadar” → “Job Descriptions”).
Credential Required: Connect your airtableTokenApi credentials.
Add Match Job From Subject and keep the provided JavaScript that computes similarity and a MIN_SCORE threshold of 0.60.
Add Match Found Check and set the boolean condition to {{ $json.matchFound }} equals true.

Match Found Check outputs to both Initial Stream Merge and Compose AI Prompt in parallel.

Step 3: Build the AI Prompt and Prepare Resume Files

Create a detailed prompt for Gemini and extract valid resume attachments.

In Compose AI Prompt, retain the prompt template that injects EMAIL and JOB sections and returns strict JSON.
Connect Initial Stream Merge to Prepare Resume Files to pass email payloads.
In Prepare Resume Files, keep the allowed file types list set to application/pdf and the binary property as data.
Note the parallel split: Prepare Resume Files outputs to Upload Resume File, Archive To Drive, and Aggregate Result Data in parallel.

Tip: If applicants send non-PDF resumes, add allowed MIME types in Prepare Resume Files to avoid dropping files.

Step 4: Upload to Gemini and Request AI Analysis

Upload the resume file, assemble the model payload, and send the analysis request.

In Upload Resume File, set URL to https://generativelanguage.googleapis.com/upload/v1beta/files?uploadType=media and Method to POST.
Set Content Type to binaryData, Input Data Field Name to data, and include headers: x-goog-api-key = {{$env.GEMINI_API_KEY}} and Content-Type = {{ $json.mimeType || $binary.data.mimeType || 'application/octet-stream' }}.
Connect Upload Resume File and Compose AI Prompt into Unify Prompt And File so both inputs are merged.
In Assemble Gemini Payload, keep the code that creates body with responseMimeType set to application/json.
In Request AI Analysis, set URL to https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash:generateContent and JSON Body to {{$json.body}}.
Set the query parameter key to {{$env.GEMINI_API_KEY}} and header Content-Type to application/json.

⚠️ Common Pitfall: Both Upload Resume File and Request AI Analysis rely on the GEMINI_API_KEY environment variable—ensure it is set in n8n before testing.

Step 5: Archive and Aggregate Results

Store the resume in Drive, combine streams, parse AI output, and prepare structured fields for your spreadsheet.

Configure Archive To Drive with your target Drive and Folder.
Set Name to the provided dynamic expression to create timestamped filenames: {{ (function () { const email = ($json.fromEmail || 'unknown').toLowerCase(); const user = email.split('@')[0].replace(/[^a-z0-9._-]+/g,'-').replace(/^-+|-+$/g,''); const d = new Date(Number($json.internalDate) || Date.now()); const yyyy=d.getUTCFullYear(), mm=String(d.getUTCMonth()+1).padStart(2,'0'), dd=String(d.getUTCDate()).padStart(2,'0'), hh=String(d.getUTCHours()).padStart(2,'0'), mi=String(d.getUTCMinutes()).padStart(2,'0'), ss=String(d.getUTCSeconds()).padStart(2,'0'); const ext = (($json.fileName || '').split('.').pop() || ($json.mimeType==='application/pdf'?'pdf':'bin')).toLowerCase(); return `${user || 'unknown'}__${yyyy}-${mm}-${dd}-${hh}${mi}${ss}.${ext}`; })() }}.
Credential Required: Connect your googleDriveOAuth2Api credentials.
Keep Aggregate Result Data set to Mode combine, Combine By combineByPosition, and Number Inputs 4.
In Parse AI Output, retain the JSON parsing logic and output fields like JobTitle, CandidateName, CandidateEmail, ResumeLink, and AIScore.

Tip: Aggregate Result Data merges four inputs, including AI, file upload, and archive metadata, so avoid changing node order unless you also update downstream parsing logic.

Step 6: Configure Output to Google Sheets

Append the structured screening results to a spreadsheet for tracking.

Add Append To Sheet with Operation set to append.
Select the target Document (e.g., “ResumeRadar - Applicants”) and Sheet (e.g., “Sheet1”).
Map columns using the provided expressions such as {{ $json.CandidateEmail }}, {{ $json.CandidatePhoneNumber }}, and {{ '=HYPERLINK("' + $json.ResumeLink + '","Download")' }}.
Credential Required: Connect your googleSheetsOAuth2Api credentials.

Step 7: Test and Activate Your Workflow

Validate end-to-end execution and switch the automation to live mode.

Click Execute Workflow and send a test email with a PDF resume attached to the monitored inbox.
Confirm that Archive To Drive creates a file in your Drive folder and Append To Sheet adds a new row with the AI score and resume link.
Check the Gemini calls in Upload Resume File and Request AI Analysis for successful 200 responses.
Turn on the workflow with the Active toggle for production use.

🔒

Unlock Full Step-by-Step Guide

Get the complete implementation guide + downloadable template

Common Gotchas

Google credentials can expire or need specific permissions. If things break, check the Gmail/Drive/Sheets credentials inside n8n’s Credentials page 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.
Gemini API access can fail due to a missing key or billing restrictions. Confirm the key is set in n8n credentials or environment variables, then verify it in your Google AI Studio project.

Frequently Asked Questions

How long does it take to set up this Gmail resume scoring automation?

About an hour if your Airtable jobs and Google Sheet are ready.

Do I need coding skills to automate Gmail resume scoring?

No. You’ll mostly connect accounts and paste in your scoring criteria. The “code” nodes are already built into the workflow.

Is n8n free to use for this Gmail resume scoring 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 costs.

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.

Can I customize this Gmail resume scoring workflow for multiple departments?

Yes. You can adjust the subject matching in “Match Job From Subject” and expand your Airtable job table with department fields. Common tweaks include separate Drive folders per department, different Gemini scoring prompts per role, and extra columns in Google Sheets for hiring manager ownership.

Why is my Gmail connection failing in this workflow?

Usually it’s expired OAuth access or missing Gmail permissions in your Google account. Reconnect the Gmail credential in n8n, then confirm the trigger is watching the right mailbox/label. If you’re using a shared inbox, make sure the account you authorized actually has access to it. Also check Google’s security alerts, because a password change can silently break older sessions.

How many resumes can this Gmail resume scoring automation handle?

A lot, as long as your n8n plan and Google/Gemini limits can keep up.

Is this Gmail resume scoring automation better than using Zapier or Make?

Often, yes, because this flow benefits from branching, merges, and handling files plus AI calls in one place. n8n also gives you a self-hosting option, which can matter when resume volume grows and per-task pricing starts to sting. Zapier or Make can still be fine if you only want “email arrives → add row,” with no job matching and no file storage discipline. The moment you need job context from Airtable and consistent scoring prompts, n8n tends to be easier to maintain long-term. Talk to an automation expert if you want help choosing.

Once this is running, your inbox stops being a hiring bottleneck. The workflow handles the repetitive sorting and scoring so you can spend your attention on actual interviews.

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": "",
"meta": {
"instanceId": "846a24a35e284da49d8495de7119dffa622eb4851554a7278d5b583ab81b93c0",
"templateId": null,
"templateCredsSetupCompleted": null
},
"name": "Automated Resume Screening Flow",
"tags": [],
"nodes": [
{
"id": "flowpast-topbar-7469",
"name": "Flowpast Branding",
"type": "n8n-nodes-base.stickyNote",
"position": [
0,
20
],
"parameters": {
"color": 7,
"width": 1288,
"height": 80,
"content": "## Flowpast.com | Automation Workflow Library\n**\ud83d\udcd6 Full tutorial & setup guide:** flowpast.com"
},
"typeVersion": 1
},
{
"id": "049242f2-4300-462a-9aa0-4b02dfde99c1",
"name": "Inbox Attachment Watch",
"type": "n8n-nodes-base.gmailTrigger",
"position": [
697,
120
],
"parameters": {
"simple": false,
"filters": {
"q": "has:attachment"
},
"options": {
"downloadAttachments": true
},
"pollTimes": {
"item": [
{
"mode": "everyMinute"
}
]
}
},
"credentials": {
"gmailOAuth2": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 1.2
},
{
"id": "c610f1ed-df34-4bec-a308-ffa68b1eda74",
"name": "Retrieve Job Listings",
"type": "n8n-nodes-base.airtable",
"position": [
394,
242
],
"parameters": {
"base": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "ResumeRadar"
},
"table": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "Job Descriptions"
},
"options": [],
"operation": "search"
},
"credentials": {
"airtableTokenApi": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 2.1
},
{
"id": "424f557a-1bab-4116-ad34-2f03251f44d9",
"name": "Match Job From Subject",
"type": "n8n-nodes-base.code",
"position": [
226,
194
],
"parameters": {
"jsCode": "// ---------- config ----------\nconst MIN_SCORE = 0.60; // raise/lower to be stricter/looser\n// ----------------------------\n\n// Build job list from Airtable items\nconst jobListings = [];\nif (Array.isArray(items)) {\n  for (const item of items) {\n    if (item?.json) {\n      jobListings.push({\n        id:          item.json.id || '',\n        title:       item.json[\"Job Title\"] || '',\n        jobCode:     item.json[\"Job Code\"] || '',\n        description: item.json[\"Description\"] || '',\n        skills:      item.json[\"Skills Required\"] || '',\n        experience:  item.json[\"Experience Required\"] || item.json[\"Experience Required (Years)\"] || '',\n        location:    item.json[\"Location\"] || ''\n      });\n    }\n  }\n}\n\n// Email subject from Gmail Trigger (Option A: Simplify OFF)\nconst emailNode = $node[\"Inbox Attachment Watch\"];\nconst email = emailNode?.json || {};\nlet subject = email.subject || email.Subject || '';\n\nif (!subject && items?.[0]?.json) {\n  subject = items[0].json.subject || items[0].json.Subject || '';\n}\n\nfunction extractJobTitle(subj) {\n  const lower = (subj || '').toLowerCase();\n  const patterns = [\n    /application (?:for|to) (.*?)(?:position|role|job)?$/i,\n    /(?:applying|apply) (?:for|to) (?:the )?(.*?)(?:position|role|job)?$/i,\n    /interested in (?:the )?(.*?)(?:position|role|job)/i,\n    /regarding (?:the )?(.*?)(?:position|role|job)/i,\n    /^(.*?)(?:position|job|role|application)$/i,\n  ];\n  for (const p of patterns) {\n    const m = lower.match(p);\n    if (m?.[1]?.trim()?.length > 3) return m[1].trim();\n  }\n  const capRe = /([A-Z][a-z]+(?: [A-Z][a-z]+)+)/g;\n  let best = '', m;\n  while ((m = capRe.exec(subj || '')) !== null) if (m[0].length > best.length) best = m[0];\n  if (best.length > 5) return best.toLowerCase();\n  return lower\n    .replace(/application for|applying for|regarding|re:|job application|position|role|job|application|interest in/g, '')\n    .replace(/^for the |for a |for |the |a /g, '')\n    .replace(/\\s+/g, ' ')\n    .trim();\n}\n\n// ---------- NEW similarity ----------\nconst STOP = new Set(['the','a','an','and','for','to','of','role','position','job','senior','jr','junior','lead','manager','developer','engineer']);\nconst PHRASE_REPL = [\n  ['machine learning','ai'],\n  ['artificial intelligence','ai'],\n  ['quality assurance','qa'],\n  ['full stack','fullstack'],\n  ['front end','frontend'],\n  ['back end','backend'],\n];\nconst TOKEN_MAP = { // single-token synonyms\n  ml: 'ai',\n  sdet: 'qa',\n  testing: 'qa',\n  test: 'qa',\n  sre: 'devops',\n};\n\nfunction normalizeTitle(s = '') {\n  s = s.toLowerCase();\n  for (const [a,b] of PHRASE_REPL) s = s.replace(new RegExp(a, 'g'), b);\n  s = s.replace(/[\\/,()\\-]/g, ' ');\n  s = s.replace(/\\s+/g, ' ').trim();\n  return s;\n}\n\nfunction canonicalTokens(s = '') {\n  const t = normalizeTitle(s).split(' ').filter(Boolean);\n  const out = [];\n  for (const tok of t) {\n    if (STOP.has(tok)) continue;\n    out.push(TOKEN_MAP[tok] || tok);\n  }\n  // dedupe, preserve order\n  return Array.from(new Set(out));\n}\n\nfunction jaccard(a, b) {\n  const A = new Set(a), B = new Set(b);\n  if (A.size === 0 && B.size === 0) return 1;\n  const inter = new Set([...A].filter(x => B.has(x))).size;\n  const uni = new Set([...A, ...B]).size;\n  return uni ? inter / uni : 0;\n}\n\nfunction lev(a, b) {\n  a = normalizeTitle(a); b = normalizeTitle(b);\n  if (!a && !b) return 1;\n  const T = Array(b.length + 1).fill(null).map(() => Array(a.length + 1).fill(null));\n  for (let i = 0; i <= a.length; i++) T[0][i] = i;\n  for (let j = 0; j <= b.length; j++) T[j][0] = j;\n  for (let j = 1; j <= b.length; j++) {\n    for (let i = 1; i <= a.length; i++) {\n      const cost = a[i - 1] === b[j - 1] ? 0 : 1;\n      T[j][i] = Math.min(T[j][i - 1] + 1, T[j - 1][i] + 1, T[j - 1][i - 1] + cost);\n    }\n  }\n  const d = T[b.length][a.length];\n  const maxLen = Math.max(a.length, b.length) || 1;\n  return 1 - d / maxLen;\n}\n\n// final score = 70% token match + 30% string similarity\nfunction smartSim(s1, s2) {\n  const jac = jaccard(canonicalTokens(s1), canonicalTokens(s2));\n  const levSim = lev(s1, s2);\n  return 0.7 * jac + 0.3 * levSim;\n}\n// ------------------------------------\n\nconst jobKeywords = extractJobTitle(subject);\n\n// guard: no keywords\nif (!jobKeywords) {\n  return [{\n    json: {\n      subject,\n      extractedJobTitle: jobKeywords,\n      matchFound: false,\n      reason: \"No extractable job title from subject\"\n    },\n  }];\n}\n\n// find best match using smartSim\nlet best = null, bestScore = 0;\nfor (const j of jobListings) {\n  const s = smartSim(jobKeywords, j.title);\n  if (s > bestScore) { best = j; bestScore = s; }\n}\n\n// Return the match\nif (best && bestScore >= MIN_SCORE) {\n  return [{\n    json: {\n      subject,\n      extractedJobTitle: jobKeywords,\n      matchFound: true,\n      jobMatch: {\n        id: best.id,\n        title: best.title,\n        code: best.jobCode,\n        description: best.description,\n        skills: best.skills,\n        experience: best.experience,\n        location: best.location,\n        confidence: bestScore\n      }\n    },\n  }];\n}\n\nreturn [{\n  json: {\n    subject,\n    extractedJobTitle: jobKeywords,\n    matchFound: false,\n    reason: best\n      ? `Low confidence (${bestScore.toFixed(2)}) vs threshold ${MIN_SCORE}`\n      : 'No jobs in list'\n  },\n}];"
},
"typeVersion": 2
},
{
"id": "2873ae38-0c44-4493-a604-1088e768ec63",
"name": "Match Found Check",
"type": "n8n-nodes-base.if",
"position": [
14,
202
],
"parameters": {
"options": [],
"conditions": {
"options": {
"version": 2,
"leftValue": "",
"caseSensitive": true,
"typeValidation": "strict"
},
"combinator": "and",
"conditions": [
{
"id": "66b8762f-c4ee-46c4-b2e5-a9e78ffccebd",
"operator": {
"type": "boolean",
"operation": "true",
"singleValue": true
},
"leftValue": "={{ $json.matchFound }}",
"rightValue": ""
}
]
}
},
"typeVersion": 2.2
},
{
"id": "a528ce65-eeb4-49c4-8c8d-bf152ea6f52a",
"name": "Compose AI Prompt",
"type": "n8n-nodes-base.code",
"position": [
0,
394
],
"parameters": {
"jsCode": "// PromptBuilder \u2014 builds the LLM prompt + carries email body & job meta\n\nconst m   = items[0]?.json ?? {};\nconst job = m.jobMatch ?? {};\n\n// helpers\nfunction stripHtml(s = '') {\n  return String(s)\n    .replace(/<style[\\s\\S]*?<\\/style>/gi, '')\n    .replace(/<script[\\s\\S]*?<\\/script>/gi, '')\n    .replace(/<[^>]+>/g, ' ')\n    .replace(/\\s+/g, ' ')\n    .trim();\n}\nfunction firstStr(...vals) {\n  for (const v of vals) if (typeof v === 'string' && v.trim()) return v.trim();\n  return '';\n}\n\n// get Gmail payload (paired item first, then pinned/previous run as fallback)\nconst g1 = $item(0, 'Inbox Attachment Watch')?.$json ?? {};\nconst g2 = (($items && $items('Inbox Attachment Watch', 0)) || [])[0]?.json ?? {};\n\n// build email body: prefer any body already carried on the item, else read from Gmail\nlet emailBody = firstStr(\n  m.emailBody,\n  g1.text, g1.textPlain,\n  g1.html ? stripHtml(g1.html) : '',\n  g1.textHtml ? stripHtml(g1.textHtml) : '',\n  g1.textAsHtml ? stripHtml(g1.textAsHtml) : '',\n  g1.snippet,\n  g2.text, g2.textPlain,\n  g2.html ? stripHtml(g2.html) : '',\n  g2.textHtml ? stripHtml(g2.textHtml) : '',\n  g2.textAsHtml ? stripHtml(g2.textAsHtml) : '',\n  g2.snippet\n);\nemailBody = emailBody.slice(0, 5000); // keep prompt compact\n\n// subject: prefer the one already on the item (from EmailMatcher), else Gmail\nconst subject = firstStr(m.subject, g1.subject, g2.subject);\n\n// prompt\nconst prompt = `\nYou are a senior technical recruiter.\n\nTASKS\n1) Read the ATTACHED resume (primary source of truth).\n2) Consider the email subject & body for extra context.\n3) Evaluate fit against the JOB description.\n\nExtract ONLY from the resume (do NOT infer from email headers):\n- candidateName: full legal name as printed in the resume. If missing, return \"\".\n- candidatePhone: best reachable phone (normalize to E.164 if possible; else raw). If missing, return \"\".\n\nScoring:\n- score: integer 1\u201310 (10=excellent, 7\u20139=strong, 4\u20136=partial, 1\u20133=poor).\n- explanation: one concise sentence under 40 words.\n\nEMAIL\n- Subject: ${subject}\n- Body: ${emailBody}\n\nJOB\n- Title: ${job.title || \"\"}\n- Code: ${job.code || \"\"}\n- Description: ${job.description || \"\"}\n- Skills Required: ${job.skills || \"\"}\n- Experience Required: ${job.experience || \"\"}\n- Location: ${job.location || \"\"}\n\nReturn STRICT JSON only (no prose, no markdown):\n{\n  \"score\": <1-10>,\n  \"explanation\": \"<under 40 words>\",\n  \"candidateName\": \"<from resume or empty string>\",\n  \"candidatePhone\": \"<E.164 if possible else raw or empty string>\"\n}\n`.trim();\n\nreturn [{ json: { prompt, jobMeta: job } }];\n"
},
"typeVersion": 2
},
{
"id": "f0c3079a-8cda-4a2c-bfde-0fc3d61037bd",
"name": "Initial Stream Merge",
"type": "n8n-nodes-base.merge",
"position": [
242,
482
],
"parameters": [],
"typeVersion": 3.2
},
{
"id": "39fc01a4-5a9b-438e-9932-e4b82f47bb4c",
"name": "Prepare Resume Files",
"type": "n8n-nodes-base.code",
"position": [
392,
450
],
"parameters": {
"jsCode": "const matched = items.some(i => i.json?.matchFound === true || i.json?.jobMatch);\nif (!matched) return [];\n\n\n// Keep only real resume-ish attachments\nconst ALLOWED = new Set([\n  'application/pdf',\n]);\n\n    \n// --- helpers ---\nfunction headerPick(g, key) {\n  const v = g.headers?.[key] || g.headers?.[key.toLowerCase()];\n  return v ? String(v).replace(/^[A-Za-z-]+:\\s*/, '') : '';\n}\nfunction pickEmail(str) {\n  const m = String(str || '').match(/[A-Z0-9._%+-]+@[A-Z0-9.-]+\\.[A-Z]{2,}/i);\n  return m ? m[0].toLowerCase() : '';\n}\nfunction pickName(str) {\n  // \"Display Name\" <email@domain>\n  const m = String(str || '').match(/^\"?([^\"<]+?)\"?\\s*<[^>]+>/);\n  return m ? m[1].trim() : '';\n}\nfunction extractSender(g) {\n  // 1) Preferred: rich shape\n  const v0 = g.from?.value?.[0];\n  const addr = v0?.address || '';\n  const nm   = v0?.name || '';\n\n  // 2) Text forms (rich shape has these too)\n  const fromText = typeof g.from?.text === 'string' ? g.from.text : '';\n  const replyTo  = headerPick(g, 'reply-to');\n  const fromHdr  = headerPick(g, 'from') || fromText;\n  const returnPath = headerPick(g, 'return-path');\n\n  // Decide email\n  const email = addr || pickEmail(fromHdr) || pickEmail(replyTo) || pickEmail(returnPath) || '';\n  // Decide name\n  let name = nm || pickName(fromHdr);\n  if (!name && email) name = email.split('@')[0];\n\n  return { email, name };\n}\n\nconst out = [];\nfor (const item of items) {\n  const g = item.json || {};\n  const { email: fromEmail, name: fromName } = extractSender(g);\n\n  // Prefer internalDate (epoch); else parse ISO \"date\"; else now\n  const when =\n    (g.internalDate ? Number(g.internalDate) : 0) ||\n    (g.date ? Date.parse(g.date) : 0) ||\n    Date.now();\n\n  if (!item.binary) continue;\n  for (const key of Object.keys(item.binary)) {\n    const bin = item.binary[key];\n    if (!bin?.data) continue;\n    if (ALLOWED.size && !ALLOWED.has(bin.mimeType)) continue;\n\n    out.push({\n      json: {\n        fileName: bin.fileName || 'resume',\n        mimeType: bin.mimeType || 'application/octet-stream',\n        fromEmail,\n        fromName,\n        internalDate: when,\n        subject: g.subject || headerPick(g, 'subject') || ''\n      },\n      binary: { data: bin } // Downstream nodes use Binary Property: \"data\"\n    });\n  }\n}\nreturn out;\n"
},
"typeVersion": 2,
"alwaysOutputData": false
},
{
"id": "c18f2954-dced-4342-a0aa-66a81545f5ff",
"name": "Upload Resume File",
"type": "n8n-nodes-base.httpRequest",
"position": [
648,
488
],
"parameters": {
"url": "https://generativelanguage.googleapis.com/upload/v1beta/files?uploadType=media",
"method": "POST",
"options": [],
"sendBody": true,
"contentType": "binaryData",
"sendHeaders": true,
"headerParameters": {
"parameters": [
{
"name": "x-goog-api-key",
"value": "={{$env.GEMINI_API_KEY}}"
},
{
"name": "Content-Type",
"value": "={{ $json.mimeType || $binary.data.mimeType || 'application/octet-stream' }}"
}
]
},
"inputDataFieldName": "data"
},
"typeVersion": 4.2
},
{
"id": "ee2d7f60-58be-475c-8d54-502f0bbf0aea",
"name": "Unify Prompt And File",
"type": "n8n-nodes-base.merge",
"position": [
778,
498
],
"parameters": [],
"typeVersion": 3.2
},
{
"id": "08d15341-fad1-41b4-8dd1-45c19ef66e19",
"name": "Assemble Gemini Payload",
"type": "n8n-nodes-base.code",
"position": [
998,
448
],
"parameters": {
"jsCode": "const inItems = items.map(i => i.json);\n\n// 1) file object from FileUpload\nconst fileBlock = inItems.find(o => o.file?.uri)?.file;\nif (!fileBlock?.uri) throw new Error('Missing Gemini file URI');\n\n// 2) prompt/meta from PromptBuilder\nconst metaBlock = inItems.find(o => typeof o.prompt === 'string') || {};\nconst jm = metaBlock.jobMeta || {};\n\n// 3) prefer the ORIGINAL attachment mimeType (from ReadyAttachment);\n//    fallback to the FileUpload's mimeType; final fallback to octet-stream.\nconst realMime =\n  (inItems.find(o => o.mimeType)?.mimeType) ||      // ReadyAttachment\n  fileBlock.mimeType ||                             // FileUpload (if correct)\n  'application/octet-stream';\n\nconst parts = [];\nif (metaBlock.prompt) parts.push({ text: metaBlock.prompt });\nif (metaBlock.emailBody) parts.push({ text: `\\nEMAIL BODY:\\n${metaBlock.emailBody}` });\n\nparts.push({\n  text:\n`JOB META:\nTitle: ${jm.title ?? ''}\nCode: ${jm.code ?? ''}\nDescription: ${jm.description ?? ''}\nSkills: ${jm.skills ?? ''}\nExperience: ${jm.experience ?? ''}\nLocation: ${jm.location ?? ''}`\n});\n\n// \u2705 pass the real mime type here\nparts.push({\n  fileData: {\n    fileUri: fileBlock.uri,\n    mimeType: realMime\n  }\n});\n\nconst body = {\n  contents: [{ role: 'user', parts }],\n  generationConfig: {\n    temperature: 0.1,\n    responseMimeType: 'application/json',\n    responseSchema: {\n      type: 'object',\n      properties: {\n        score: { type: 'integer', minimum: 1, maximum: 10 },\n        explanation: { type: 'string' },\n        candidateName: { type: 'string' },\n        candidatePhone: { type: 'string' }\n      },\n      required: ['score','explanation']\n    },\n  },\n};\n\nreturn [{ json: { body } }];\n"
},
"typeVersion": 2
},
{
"id": "14be41de-3ef7-4258-952d-affe1c439341",
"name": "Request AI Analysis",
"type": "n8n-nodes-base.httpRequest",
"position": [
1102,
500
],
"parameters": {
"url": "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash:generateContent",
"method": "POST",
"options": [],
"jsonBody": "={{$json.body}}",
"sendBody": true,
"sendQuery": true,
"sendHeaders": true,
"specifyBody": "json",
"queryParameters": {
"parameters": [
{
"name": "key",
"value": "={{$env.GEMINI_API_KEY}}"
}
]
},
"headerParameters": {
"parameters": [
{
"name": "Content-Type",
"value": "application/json"
}
]
}
},
"typeVersion": 4.2
},
{
"id": "2ef593bc-bd85-451b-91a4-e93c32739659",
"name": "Aggregate Result Data",
"type": "n8n-nodes-base.merge",
"position": [
890,
626
],
"parameters": {
"mode": "combine",
"options": [],
"combineBy": "combineByPosition",
"numberInputs": 4
},
"typeVersion": 3.2
},
{
"id": "03a33b24-68a2-4f5f-afc2-36ef1cbdb0f7",
"name": "Archive To Drive",
"type": "n8n-nodes-base.googleDrive",
"position": [
596,
632
],
"parameters": {
"name": "={{\n  (function () {\n    const email = ($json.fromEmail || 'unknown').toLowerCase();\n    const user = email.split('@')[0].replace(/[^a-z0-9._-]+/g,'-').replace(/^-+|-+$/g,'');\n    const d = new Date(Number($json.internalDate) || Date.now());\n    const yyyy=d.getUTCFullYear(), mm=String(d.getUTCMonth()+1).padStart(2,'0'),\n          dd=String(d.getUTCDate()).padStart(2,'0'),\n          hh=String(d.getUTCHours()).padStart(2,'0'),\n          mi=String(d.getUTCMinutes()).padStart(2,'0'),\n          ss=String(d.getUTCSeconds()).padStart(2,'0');\n    const ext = (($json.fileName || '').split('.').pop()\n              || ($json.mimeType==='application/pdf'?'pdf':'bin')).toLowerCase();\n    return `${user || 'unknown'}__${yyyy}-${mm}-${dd}-${hh}${mi}${ss}.${ext}`;\n  })()\n}}\n",
"driveId": {
"__rl": true,
"mode": "list",
"value": "My Drive"
},
"options": [],
"folderId": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "ResumeRadar"
}
},
"credentials": {
"googleDriveOAuth2Api": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 3
},
{
"id": "e0dd0699-059a-49ba-942d-58f10aaec4b7",
"name": "Parse AI Output",
"type": "n8n-nodes-base.code",
"position": [
1064,
676
],
"parameters": {
"jsCode": "// Input is the ONE merged item\nconst j = items[0].json;\n\n// Parse model JSON\nconst text = j?.candidates?.[0]?.content?.parts?.[0]?.text ?? '{}';\nlet data = {};\ntry { data = JSON.parse(text); } catch {}\n\nconst normPhone = s => s ? s.toString().trim().replace(/[^\\d+]/g,'') : '';\n\n// Build the exact columns we want for Sheets\nreturn [{\n  json: {\n    JobTitle: j.jobMeta?.title || j.jobMatch?.title || j.subject || '',\n    CandidateName: (data.candidateName || '').trim() || (j.fromName || ''),\n    CandidateEmail: j.fromEmail || '',\n    CandidatePhoneNumber: normPhone(data.candidatePhone || ''),\n    // prefer open link; fallback to direct or ID\n    ResumeLink: j.webViewLink || j.webContentLink || (j.id ? `https://drive.google.com/file/d/${j.id}/view?usp=drivesdk` : ''),\n    AIScore: data.score ?? null,\n    AIExplanation: data.explanation ?? ''\n  }\n}];\n"
},
"typeVersion": 2
},
{
"id": "3cf0a49e-df9a-4cca-b245-a46838748e69",
"name": "Append To Sheet",
"type": "n8n-nodes-base.googleSheets",
"position": [
1188,
622
],
"parameters": {
"columns": {
"value": {
"Email": "={{ $json.CandidateEmail }}",
"Resume": "={{ '=HYPERLINK(\"' + $json.ResumeLink + '\",\"Download\")' }}",
"Contact": "={{ $json.CandidatePhoneNumber }}",
"AI Score": "={{ $json.AIScore }}",
"Job Title": "={{ $json.JobTitle }}",
"AI Explanation": "={{ $json.AIExplanation }}",
"Candidate Name": "={{ $json.CandidateName }}"
},
"schema": [
{
"id": "Job Title",
"type": "string",
"display": true,
"required": false,
"displayName": "Job Title",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "Candidate Name",
"type": "string",
"display": true,
"required": false,
"displayName": "Candidate Name",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "Email",
"type": "string",
"display": true,
"required": false,
"displayName": "Email",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "Contact",
"type": "string",
"display": true,
"required": false,
"displayName": "Contact",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "Resume",
"type": "string",
"display": true,
"required": false,
"displayName": "Resume",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "AI Score",
"type": "string",
"display": true,
"required": false,
"displayName": "AI Score",
"defaultMatch": false,
"canBeUsedToMatch": true
},
{
"id": "AI Explanation",
"type": "string",
"display": true,
"required": false,
"displayName": "AI Explanation",
"defaultMatch": false,
"canBeUsedToMatch": true
}
],
"mappingMode": "defineBelow",
"matchingColumns": [],
"attemptToConvertTypes": false,
"convertFieldsToString": false
},
"options": [],
"operation": "append",
"sheetName": {
"__rl": true,
"mode": "list",
"value": "gid=0",
"cachedResultUrl": "",
"cachedResultName": "Sheet1"
},
"documentId": {
"__rl": true,
"mode": "list",
"value": "[YOUR_ID]",
"cachedResultUrl": "",
"cachedResultName": "ResumeRadar - Applicants"
}
},
"credentials": {
"googleSheetsOAuth2Api": {
"id": "credential-id",
"name": ""
}
},
"typeVersion": 4.6
}
],
"active": true,
"pinData": [],
"settings": {
"callerPolicy": "workflowsFromSameOwner",
"executionOrder": "v1",
"saveExecutionProgress": true
},
"versionId": "",
"connections": {
"Match Found Check": {
"main": [
[
{
"node": "Initial Stream Merge",
"type": "main",
"index": 1
},
{
"node": "Compose AI Prompt",
"type": "main",
"index": 0
}
],
[]
]
},
"Unify Prompt And File": {
"main": [
[
{
"node": "Assemble Gemini Payload",
"type": "main",
"index": 0
}
]
]
},
"Initial Stream Merge": {
"main": [
[
{
"node": "Prepare Resume Files",
"type": "main",
"index": 0
}
]
]
},
"Retrieve Job Listings": {
"main": [
[
{
"node": "Match Job From Subject",
"type": "main",
"index": 0
}
]
]
},
"Upload Resume File": {
"main": [
[
{
"node": "Unify Prompt And File",
"type": "main",
"index": 0
}
]
]
},
"Aggregate Result Data": {
"main": [
[
{
"node": "Parse AI Output",
"type": "main",
"index": 0
}
]
]
},
"Match Job From Subject": {
"main": [
[
{
"node": "Match Found Check",
"type": "main",
"index": 0
}
]
]
},
"Archive To Drive": {
"main": [
[
{
"node": "Aggregate Result Data",
"type": "main",
"index": 0
}
]
]
},
"Inbox Attachment Watch": {
"main": [
[
{
"node": "Retrieve Job Listings",
"type": "main",
"index": 0
},
{
"node": "Initial Stream Merge",
"type": "main",
"index": 0
}
]
]
},
"Parse AI Output": {
"main": [
[
{
"node": "Append To Sheet",
"type": "main",
"index": 0
}
]
]
},
"Compose AI Prompt": {
"main": [
[
{
"node": "Unify Prompt And File",
"type": "main",
"index": 1
},
{
"node": "Aggregate Result Data",
"type": "main",
"index": 3
}
]
]
},
"Assemble Gemini Payload": {
"main": [
[
{
"node": "Request AI Analysis",
"type": "main",
"index": 0
}
]
]
},
"Prepare Resume Files": {
"main": [
[
{
"node": "Upload Resume File",
"type": "main",
"index": 0
},
{
"node": "Archive To Drive",
"type": "main",
"index": 0
},
{
"node": "Aggregate Result Data",
"type": "main",
"index": 1
}
]
]
},
"Request AI Analysis": {
"main": [
[
{
"node": "Aggregate Result Data",
"type": "main",
"index": 2
}
]
]
}
}
}