This workflow follows the Gmail → Gmail Trigger recipe pattern — see all workflows that pair these two integrations.
The workflow JSON
Copy or download the full n8n JSON below. Paste it into a new n8n workflow, add your credentials, activate. Full import guide →
{
"name": "Reference Check Parser \u2013 Turn Reference Emails into a Comparable Candidate Sheet (powered by easybits)",
"nodes": [
{
"parameters": {
"pollTimes": {
"item": [
{
"mode": "everyMinute"
}
]
},
"simple": false,
"filters": {
"labelIds": []
},
"options": {
"downloadAttachments": true
}
},
"type": "n8n-nodes-base.gmailTrigger",
"typeVersion": 1.3,
"position": [
0,
0
],
"id": "c10f60e9-13b4-4d44-a3ff-be8f1304ff72",
"name": "Gmail Trigger",
"credentials": {
"gmailOAuth2": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"jsCode": "const item = $input.first();\nconst binary = item.binary || {};\nconst out = [];\n\nfor (const key of Object.keys(binary)) {\n const meta = binary[key];\n out.push({\n json: {\n file_name: meta.fileName || '',\n mime_type: meta.mimeType || '',\n source_subject: item.json.subject || '',\n source_from: item.json.from || '',\n },\n binary: { data: meta }, // normalize every attachment to the key \"data\"\n });\n}\n\nreturn out;"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
336,
0
],
"id": "efdf057b-3929-401c-9bfe-a6b67fe18372",
"name": "Split Out: Attachments"
},
{
"parameters": {
"conditions": {
"options": {
"caseSensitive": true,
"leftValue": "",
"typeValidation": "strict",
"version": 3
},
"conditions": [
{
"id": "fbb5271c-9f22-486b-9489-8c08f1d1b7cc",
"leftValue": "={{ [\"application/pdf\",\"image/jpeg\",\"image/png\",\"image/tiff\"].includes($json.mime_type) }}",
"rightValue": "",
"operator": {
"type": "boolean",
"operation": "true",
"singleValue": true
}
},
{
"id": "7de102d9-bfad-4d23-90e3-b54e88378740",
"leftValue": "={{ !/logo|signature|footer|icon|banner/i.test($json.file_name) }}",
"rightValue": "",
"operator": {
"type": "boolean",
"operation": "true",
"singleValue": true
}
}
],
"combinator": "and"
},
"options": {}
},
"type": "n8n-nodes-base.filter",
"typeVersion": 2.3,
"position": [
640,
0
],
"id": "2d21dc86-f3ed-498a-ba59-3403d4269c88",
"name": "Filter: Real Documents"
},
{
"parameters": {
"options": {}
},
"type": "n8n-nodes-base.splitInBatches",
"typeVersion": 3,
"position": [
960,
0
],
"id": "c6dfac5f-b45c-45e2-8644-5f782766c685",
"name": "Loop Over Letters"
},
{
"parameters": {},
"type": "@easybits/n8n-nodes-extractor.easybitsExtractor",
"typeVersion": 2,
"position": [
1296,
176
],
"id": "e3ffd559-2671-44b0-b8cc-9cc646781216",
"name": "Extract Reference: Letter",
"credentials": {
"easybitsExtractorApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"jsCode": "const d = $input.first().json.data;\n\n// Extractor returns null for missing fields \u2014 but also catch the string \"null\" and empties\nconst isMissing = (v) =>\n v === null || v === undefined || String(v).trim().toLowerCase() === 'null' || String(v).trim() === '';\n\n// Arrays sometimes come back JSON-encoded or comma-joined\nconst toArray = (v) => {\n if (isMissing(v)) return [];\n if (Array.isArray(v)) return v;\n const s = String(v).trim();\n if (s.startsWith('[')) {\n try { return JSON.parse(s); } catch (e) { /* fall through */ }\n }\n return s.split(',').map(x => x.trim()).filter(Boolean);\n};\n\nconst clean = (v, fallback = '') => isMissing(v) ? fallback : String(v).trim();\n\n// Constrain free-text enums to known values\nconst normEnum = (v, allowed, fallback) => {\n const s = clean(v).toLowerCase();\n return allowed.includes(s) ? s : fallback;\n};\n\nconst strengths = toArray(d.claimed_strengths);\nconst weaknesses = toArray(d.claimed_weaknesses);\n\nreturn [{\n json: {\n referee_name: clean(d.referee_name, 'Unknown'),\n referee_title_company: clean(d.referee_title_company),\n candidate_name: clean(d.candidate_name, 'Unknown'),\n relationship_type: clean(d.relationship_type).toLowerCase(),\n duration_known: clean(d.duration_known),\n claimed_strengths: strengths,\n claimed_weaknesses: weaknesses,\n strengths_text: strengths.join('; '),\n weaknesses_text: weaknesses.length ? weaknesses.join('; ') : '\u2014 none stated \u2014',\n would_rehire: normEnum(d.would_rehire, ['yes', 'no', 'unstated'], 'unstated'),\n tone: normEnum(d.tone, ['positive', 'neutral', 'hedged'], 'neutral'),\n notable_quote: clean(d.notable_quote),\n }\n}];"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
1600,
176
],
"id": "e3b79213-3c0f-445f-a6cc-399691c5d308",
"name": "Normalize Reference: Fields"
},
{
"parameters": {
"jsCode": "const r = $input.first().json;\n\nlet score = 50; // neutral baseline\n\n// Tone (LLM's first-pass read, used as input, not as final grade)\nif (r.tone === 'positive') score += 25;\nif (r.tone === 'hedged') score -= 25;\n\n// Would-rehire is the single strongest signal in reference checking\nif (r.would_rehire === 'yes') score += 20;\nif (r.would_rehire === 'no') score -= 35;\n\n// Substance: a real reference names specifics\nscore += Math.min(r.claimed_strengths.length * 3, 12);\n\n// A reference with zero stated weaknesses is often a polite non-endorsement\nif (r.claimed_weaknesses.length === 0 && r.would_rehire !== 'yes') score -= 8;\n\nscore = Math.max(0, Math.min(100, score));\n\n// Tier for routing / colour-coding the sheet\nlet tier = 'mid';\nif (score >= 75) tier = 'high';\nif (score < 45) tier = 'low';\n\nreturn [{ json: { ...r, sentiment_score: score, sentiment_tier: tier } }];"
},
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
1936,
176
],
"id": "fc583421-66fe-4dcd-aead-3910abd2878c",
"name": "Score Reference: Sentiment"
},
{
"parameters": {
"operation": "append",
"documentId": {
"__rl": true,
"value": "",
"mode": "list"
},
"sheetName": {
"__rl": true,
"value": "",
"mode": "list"
},
"columns": {
"mappingMode": "defineBelow",
"value": {
"candidate name": "={{ $json.candidate_name }}",
"referee name": "={{ $json.referee_name }}",
"referee title & company": "={{ $json.referee_title_company }}",
"relationship type": "={{ $json.relationship_type }}",
"duration": "={{ $json.duration_known }}",
"tone": "={{ $json.tone }}",
"would rehire": "={{ $json.would_rehire }}",
"sentiment score": "={{ $json.sentiment_score }}",
"sentiment tier": "={{ $json.sentiment_tier }}",
"strengths text": "={{ $json.strengths_text }}",
"weaknesses text": "={{ $json.weaknesses_text }}",
"notable quote": "={{ $json.notable_quote }}",
"received at": "={{ $now }}"
},
"matchingColumns": [],
"schema": [
{
"id": "candidate name",
"displayName": "candidate name",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "referee name",
"displayName": "referee name",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "referee title & company",
"displayName": "referee title & company",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "relationship type",
"displayName": "relationship type",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "duration",
"displayName": "duration",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "tone",
"displayName": "tone",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "would rehire",
"displayName": "would rehire",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "sentiment score",
"displayName": "sentiment score",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "sentiment tier",
"displayName": "sentiment tier",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "strengths text",
"displayName": "strengths text",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "weaknesses text",
"displayName": "weaknesses text",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "notable quote",
"displayName": "notable quote",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
},
{
"id": "received at",
"displayName": "received at",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true,
"removed": false
}
],
"attemptToConvertTypes": false,
"convertFieldsToString": false
},
"options": {}
},
"type": "n8n-nodes-base.googleSheets",
"typeVersion": 4.7,
"position": [
2256,
176
],
"id": "74f112d8-4d8b-4f13-a4e9-d895ce328d8a",
"name": "Append Row: References",
"credentials": {
"googleSheetsOAuth2Api": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"operation": "removeLabels",
"messageId": "={{ $('Gmail Trigger').item.json.id }}",
"labelIds": []
},
"type": "n8n-nodes-base.gmail",
"typeVersion": 2.2,
"position": [
1280,
-224
],
"id": "d7be8767-754b-4ff1-923a-b4cb95846cba",
"name": "Remove Label: References",
"credentials": {
"gmailOAuth2": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"operation": "addLabels",
"messageId": "={{ $('Gmail Trigger').item.json.id }}",
"labelIds": []
},
"type": "n8n-nodes-base.gmail",
"typeVersion": 2.2,
"position": [
1600,
-224
],
"id": "eeea6f5b-4682-4f05-91bc-5e573e59b339",
"name": "Add Label: Processed",
"credentials": {
"gmailOAuth2": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"content": "## \ud83d\udce5 Watch for References\nFires when an email lands under the `References` label, with **Download Attachments** on so the scanned letters come through as binary. Keep **Simplify** off \u2013 simplified output strips attachments.",
"height": 384,
"width": 304,
"color": 7
},
"type": "n8n-nodes-base.stickyNote",
"position": [
-96,
-224
],
"typeVersion": 1,
"id": "94cb8ca5-b13e-48d1-a46b-1b681fd1af1e",
"name": "Sticky Note"
},
{
"parameters": {
"content": "## \ud83d\udcce Split Out: Attachments\nTurns one email with several attachments into one item per file, since n8n stacks them on a single item. Renames every binary to the key `data` so the Extractor always reads from the same place.",
"height": 384,
"width": 304,
"color": 7
},
"type": "n8n-nodes-base.stickyNote",
"position": [
224,
-224
],
"typeVersion": 1,
"id": "3525ea95-fdaf-4bb9-8cc0-f53ee832cf6f",
"name": "Sticky Note1"
},
{
"parameters": {
"content": "## \ud83d\udee1\ufe0f Filter: Real Documents\nKeeps only real PDFs and full-page scans, dropping inline logos, signature snippets, and footer images. Guards on both file type and filename so junk attachments never reach the Extractor.",
"height": 384,
"width": 304,
"color": 7
},
"type": "n8n-nodes-base.stickyNote",
"position": [
544,
-224
],
"typeVersion": 1,
"id": "2d8e7ebd-6c41-48bb-b4af-a7095ce85092",
"name": "Sticky Note2"
},
{
"parameters": {
"content": "## \ud83d\udd01 Loop Over Letters\nProcesses one letter at a time so each becomes its own row. The **done** branch (top) runs once at the end to relabel the email; the **loop** branch (bottom) handles each letter.",
"height": 384,
"width": 304,
"color": 7
},
"type": "n8n-nodes-base.stickyNote",
"position": [
864,
-224
],
"typeVersion": 1,
"id": "76f55168-7806-4d23-92ec-56f66bed82c7",
"name": "Sticky Note3"
},
{
"parameters": {
"content": "## \ud83e\udd16 Extract Reference: Letter\nSends the letter to the **easybits Extractor**, reading binary directly from the `data` property. Pulls 10 structured fields; anything not found comes back as `null`.",
"height": 384,
"width": 304,
"color": 7
},
"type": "n8n-nodes-base.stickyNote",
"position": [
1184,
-32
],
"typeVersion": 1,
"id": "f0b0ad15-053c-438c-8b87-11dfcb57b970",
"name": "Sticky Note4"
},
{
"parameters": {
"content": "## \ud83e\uddf9 Normalize Reference: Fields\nCleans the raw output: arrays parsed safely, missing values caught, and free text snapped to fixed enums (`tone`, `would_rehire`). This is what makes references comparable across a whole column.",
"height": 384,
"width": 304,
"color": 7
},
"type": "n8n-nodes-base.stickyNote",
"position": [
1504,
-32
],
"typeVersion": 1,
"id": "b264ee3e-c3b2-4aae-bff4-5774cfcd9c5d",
"name": "Sticky Note5"
},
{
"parameters": {
"content": "## \ud83d\udcca Score Reference: Sentiment\nComputes a 0\u2013100 sentiment score in plain JavaScript \u2013 not the LLM grading itself. Would-rehire and tone carry the most weight, and a `high`/`mid`/`low` tier lets you sort the sheet at a glance.",
"height": 384,
"width": 304,
"color": 7
},
"type": "n8n-nodes-base.stickyNote",
"position": [
1824,
-32
],
"typeVersion": 1,
"id": "5ba9b7f4-4fe2-4967-b793-6a202d18c2b5",
"name": "Sticky Note6"
},
{
"parameters": {
"content": "## \ud83d\udcd7 Append Row: References\nWrites one row per letter to the `References` tab. Sorting by candidate then score stacks every reference for a person best-to-worst \u2013 the whole point of the build.",
"height": 384,
"width": 304,
"color": 7
},
"type": "n8n-nodes-base.stickyNote",
"position": [
2144,
-32
],
"typeVersion": 1,
"id": "227f72f5-d47e-4d69-86ef-f0762ecaaf1b",
"name": "Sticky Note7"
},
{
"parameters": {
"content": "## \ud83c\udff7\ufe0f Remove Label: References\nStrips the `References` label from the original email using the message ID from the trigger. This takes the email out of the trigger's scope.",
"height": 384,
"width": 304,
"color": 7
},
"type": "n8n-nodes-base.stickyNote",
"position": [
1184,
-432
],
"typeVersion": 1,
"id": "065d6a86-f6a2-4e01-80ae-1cfd57f6450d",
"name": "Sticky Note8"
},
{
"parameters": {
"content": "## \u2714\ufe0f Add Label: Processed\nTags the email with `Reference processed` so you have a clear audit trail of what's been handled.",
"height": 384,
"width": 304,
"color": 7
},
"type": "n8n-nodes-base.stickyNote",
"position": [
1504,
-432
],
"typeVersion": 1,
"id": "9a31efc9-058c-4f35-a5da-94ae230f31fd",
"name": "Sticky Note9"
},
{
"parameters": {
"content": "# \ud83d\udccb Reference Check Parser (powered by easybits)\n\nReference letters arrive as scanned attachments \u2013 multi-paragraph prose, no structure, impossible to compare. This workflow reads every letter, pulls the same 10 fields from each, scores the sentiment, and drops one comparable row per reference into a Google Sheet. Sort by candidate then score and a whole inbox of references becomes a ranked, scannable table.\n\n## How It Works\n1. **Watch** \u2013 Fires on emails under the `References` label, downloading attachments.\n2. **Split & Guard** \u2013 Splits multi-attachment emails into one item each, then filters out logos and footers.\n3. **Extract** \u2013 easybits Extractor reads each letter and returns 10 structured fields.\n4. **Normalize & Score** \u2013 Fields are cleaned, enums fixed, and a deterministic 0\u2013100 sentiment score is computed.\n5. **Log** \u2013 One row per reference lands in the `References` tab.\n6. **Relabel** \u2013 The email is marked `Reference processed` so it's never handled twice.\n\n## Setup Guide\n\n1. **easybits Extractor** \u2013 on n8n Cloud it's built in; self-hosted, install `@easybits/n8n-nodes-extractor` via Settings \u2192 Community Nodes. Free API key at easybits.tech.\n2. **Configure the 10 Extractor fields** \u2013 set them up in your easybits Extractor with the field names listed in the **Extractor Fields** section below.\n3. **Gmail** \u2013 connect your account; create a `References` label and a `Reference processed` label. Add a filter that auto-applies `References` to incoming reference emails.\n4. **Google Sheet** \u2013 one tab named `References` with these headers: `candidate_name`, `referee_name`, `referee_title_company`, `relationship_type`, `duration_known`, `tone`, `would_rehire`, `sentiment_score`, `sentiment_tier`, `strengths_text`, `weaknesses_text`, `notable_quote`, `received_at`.\n5. **Connect credentials** \u2013 Gmail (trigger + both relabel nodes), easybits, Google Sheets.\n6. **Activate** and label a reference email to test.\n\n## Extractor Fields\n- **referee_name** *(string)* \u2013 who wrote the reference\n- **referee_title_company** *(string)* \u2013 their title and company\n- **candidate_name** *(string)* \u2013 who the reference is about\n- **relationship_type** *(string)* \u2013 manager, peer, report, client, mentor\n- **duration_known** *(string)* \u2013 how long they worked together\n- **claimed_strengths** *(array)* \u2013 strengths, as short noun phrases\n- **claimed_weaknesses** *(array)* \u2013 weaknesses or growth areas\n- **would_rehire** *(string)* \u2013 yes, no, unstated\n- **tone** *(string)* \u2013 positive, neutral, hedged\n- **notable_quote** *(string)* \u2013 one standout sentence, verbatim\n\n## Notes\n- **Free plan friendly.** 10 fields fits the easybits free plan.\n- **Deterministic scoring.** The sentiment score is JavaScript, not an LLM grading its own work \u2013 would-rehire and tone weigh heaviest.\n- **No double-processing.** Relabelling drops handled emails out of the trigger's filter automatically.",
"height": 1232,
"width": 688
},
"type": "n8n-nodes-base.stickyNote",
"position": [
-800,
-608
],
"typeVersion": 1,
"id": "6107689c-2a59-47df-a305-23b433fe73b2",
"name": "Sticky Note10"
}
],
"connections": {
"Gmail Trigger": {
"main": [
[
{
"node": "Split Out: Attachments",
"type": "main",
"index": 0
}
]
]
},
"Split Out: Attachments": {
"main": [
[
{
"node": "Filter: Real Documents",
"type": "main",
"index": 0
}
]
]
},
"Filter: Real Documents": {
"main": [
[
{
"node": "Loop Over Letters",
"type": "main",
"index": 0
}
]
]
},
"Loop Over Letters": {
"main": [
[
{
"node": "Remove Label: References",
"type": "main",
"index": 0
}
],
[
{
"node": "Extract Reference: Letter",
"type": "main",
"index": 0
}
]
]
},
"Extract Reference: Letter": {
"main": [
[
{
"node": "Normalize Reference: Fields",
"type": "main",
"index": 0
}
]
]
},
"Normalize Reference: Fields": {
"main": [
[
{
"node": "Score Reference: Sentiment",
"type": "main",
"index": 0
}
]
]
},
"Score Reference: Sentiment": {
"main": [
[
{
"node": "Append Row: References",
"type": "main",
"index": 0
}
]
]
},
"Append Row: References": {
"main": [
[
{
"node": "Loop Over Letters",
"type": "main",
"index": 0
}
]
]
},
"Remove Label: References": {
"main": [
[
{
"node": "Add Label: Processed",
"type": "main",
"index": 0
}
]
]
}
},
"active": false,
"settings": {
"executionOrder": "v1",
"availableInMCP": false
},
"versionId": "",
"meta": {
"templateCredsSetupCompleted": false
},
"id": "",
"tags": []
}
Credentials you'll need
Each integration node will prompt for credentials when you import. We strip credential IDs before publishing — you'll add your own.
easybitsExtractorApigmailOAuth2googleSheetsOAuth2Api
Pro
For the full experience including quality scoring and batch install features for each workflow upgrade to Pro
About this workflow
Reference Check Parser – Turn Reference Emails into a Comparable Candidate Sheet (powered by easybits). Uses gmailTrigger, @easybits/n8n-nodes-extractor, googleSheets, gmail. Event-driven trigger; 21 nodes.
Source: https://github.com/felix-sattler-easybits/n8n-workflows/blob/850c7798a6f059900998c20ead0d7087750c17dc/easybits-candidate-reference-check-parser-workflow/easybits_candidate_reference_check_parser.json — original creator credit. Request a take-down →
Related workflows
Workflows that share integrations, category, or trigger type with this one. All free to copy and import.
AICARE Email Blast System. Uses googleDrive, httpRequest, googleSheets, gmail. Event-driven trigger; 39 nodes.
Google Drive, HTTP Request, Google Sheets +2
An automated n8n workflow that monitors your Gmail inbox, classifies job application emails using a local AI (Ollama), and logs every application — with company, role, and status — to a Google Sheet i
Gmail, Gmail Trigger, HTTP Request +1
Googlesheets Gmail. Uses googleDrive, gmailTrigger, gmail, stickyNote. Event-driven trigger; 19 nodes.
Google Drive, Gmail Trigger, Gmail +1
Splitout Code. Uses stickyNote, googleSheets, gmailTrigger, gmail. Event-driven trigger; 18 nodes.
Google Sheets, Gmail Trigger, Gmail