Extract Indian Vehicle RC Data with n8n: Drive to Fleet Sheets

Fleet ops still retype Registration Certificate (RC) fields into spreadsheets. This n8n workflow watches a Drive folder for Indian RC front and/or back images, calls Cryvis POST /v1/documents/indian-vehicle-rc, and appends structured vehicle rows to Google Sheets for fleet inventory.

Related: Automate KYC, Convert PDFs to JSON. Hub: /blog/n8n.

Architecture

Google Drive
  /fleet-rc/inbox/
    KA01AB1234_front.jpg
    KA01AB1234_back.jpg
        |
        v
n8n: pair front/back by prefix
        |
        v
POST https://api.cryvis.com/v1/documents/indian-vehicle-rc
  -F front=@...   ($binary.front)
  -F back=@...    ($binary.back)
  -H "Authorization: Bearer ..."
        |
        v
Google Sheets: Fleet registry

API contract

cURL:

curl -X POST https://api.cryvis.com/v1/documents/indian-vehicle-rc \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "front=@rc-front.jpg" \
  -F "back=@rc-back.jpg"

Docs: /docs/api/extractIndianVehicleRc · /apis/indian-vehicle-rc.

Step 1: Drive naming convention

{registration_guess}_front.jpg
{registration_guess}_back.jpg

Examples: KA01AB1234_front.jpg, KA01AB1234_back.jpg. Registration guess is only for pairing — trust Cryvis registration_number in the Sheet.

Folders:

/fleet-rc/inbox
/fleet-rc/done
/fleet-rc/failed

Step 2: Watch and pair

Google Drive Trigger on inbox.

Set (or Code) to derive pair key and side:

pair_key = {{ $json.name.replace(/_front|_back/gi, '').replace(/\.[^.]+$/, '') }}
side     = {{ $json.name.toLowerCase().includes('_front') ? 'front' : 'back' }}

When a pair_key has front only, back only, or both, proceed — Cryvis accepts any non-empty combination of front/back. Waiting for both improves completeness (source often becomes front_and_back).

Pairing strategies in n8n

1. Static Data / Data Store — on first side, write pair_key → { frontId, backId }. On second side, merge and only then call Cryvis; clear the entry after success.
2. Wait — after seeing one side, Wait 60–120s, re-list folder for the sibling file, then proceed with whatever exists (front-only still valid).
3. Manual batch — ops drops both files, then moves a READY marker that triggers the workflow.

Prefer (1) for production fleets; (3) for low volume.

Step 3: Download binaries into named properties

Download available sides. Assign binary property names explicitly:

If Drive Download always writes $binary.data, use a Code or Move Binary Data step to rename:

const items = [];
const item = $input.first();
const side = item.json.side; // 'front' | 'back'
const binary = {};
binary[side] = item.binary.data;
items.push({ json: item.json, binary });
return items;

When both sides are ready, Merge the two items into one item that carries both binary keys before HTTP.

Step 4: HTTP Request

HTTP Request → multipart:

Use IF branches or two HTTP node variants if your n8n version cannot conditionally omit multipart parts — never send an empty file part.

Response data example:

{
  "success": true,
  "data": {
    "registration_number": "KA01AB1234",
    "registration_date": "2019-06-15",
    "owner_name": "RAHUL SHARMA",
    "address": "12 MG Road, Bengaluru, Karnataka 560001",
    "vehicle_class": "LMV",
    "manufacturer": "MARUTI SUZUKI",
    "model": "SWIFT",
    "chassis_number": "MA3EJLF1S00123456",
    "engine_number": "K12MN1234567",
    "fuel_type": "Petrol",
    "insurance_valid_upto": "2025-06-14",
    "fitness_valid_upto": "2026-06-14",
    "financier": "HDFC BANK LTD",
    "source": "front_and_back"
  }
}

Map with expressions:

{{ $json.data.registration_number }}
{{ $json.data.chassis_number }}
{{ $json.data.engine_number }}
{{ $json.data.insurance_valid_upto }}
{{ $json.data.source }}

source tells you whether extraction used front, back, or both — useful when chassis is missing (often on the side you did not upload).

Step 5: Fleet Sheet columns

Google Sheets → Append Row. Optionally look up registration_number first and Update instead of Append to keep one row per vehicle.

Steps 6–8: Completeness, move, errors

IF empty chassis_number or engine_number → status_note=incomplete_identifiers (+ optional Slack). Still write the Sheet row. Success → done/{pair_key}/; error → failed/. Preserve Drive file ids with Set + Merge before HTTP.

Error Trigger: error row + Slack with pair_key.

Upsert and credits

Search Sheets on normalized registration_number (toUpperCase().replace(/\s+/g, '')): Update if found, Append if new. 1 credit per image (front+back = two). Prefer source=front_and_back. Never send file or first_page — only front and/or back.

Related reading

• Extract PAN card data
• Extract driver's license data
• Automate KYC
• Convert PDFs to JSON
• /blog/n8n

CTA

Extract Indian RC front/back to JSON: /apis/indian-vehicle-rc. Docs: /docs/api/extractIndianVehicleRc. In n8n, POST multipart front and/or back from $binary to https://api.cryvis.com/v1/documents/indian-vehicle-rc with Bearer auth and write {{ $json.data.registration_number }}, {{ $json.data.chassis_number }}, and insurance dates into your fleet Sheet.