n8n Webhook + Gmail PDF to Airtable & HubSpot Sync

Webhook + Gmail PDF to Airtable & HubSpot Sync—deduplicated, validated, and fully audited

This n8n workflow automatically ingests business records from an authenticated POST webhook or from PDF attachments in Gmail, then deduplicates, validates, and syncs updates into Airtable and HubSpot—with audit logs in Google Sheets and real-time Telegram alerts for exceptions.

What this workflow does

• Triggers from:
  • a POST webhook,
  • a Gmail label polling setup that downloads PDF attachments,
  • or a manual test run.
• Extracts structured data from PDFs in Gmail using OpenAI (LangChain Information Extractor), producing a structured batch of records.
• Splits into individual records, removes duplicates using external_id, normalizes fields, and quarantines invalid records instead of writing them.
• Prevents unnecessary updates by computing a SHA-256 content hash per record, matching Airtable by external_id, and skipping records whose content hasn’t changed.
• Upserts into Airtable, then checks HubSpot for email opt-out before upserting contacts—suppressed contacts are not synced.
• Audits every outcome: appends one audit row per record to the Google Sheets “Audit Log”, posts Telegram alerts for any non-success status, and writes a per-run summary to the “Reconcile Report”.
• Runs a daily reconciliation that counts Airtable records and the first 100 HubSpot contacts, logs counts to Google Sheets, and sends a Telegram reconcile message (silent when in sync).

Use cases

• Sync customer or lead records submitted via API into Airtable and HubSpot without duplicates.
• Extract data from inbound Gmail PDFs (e.g., forms, statements) and keep CRMs up to date automatically.
• Ensure compliance workflows by respecting HubSpot email opt-out/suppression and quarantining invalid records.

Technical details (n8n & integrations)

• Triggers: POST webhook, Gmail label polling, manual test.
• AI extraction: OpenAI via LangChain Information Extractor.
• Core logic: if, set, code, record splitting, deduping by external_id, SHA-256 hashing, and upsert logic.
• Apps: Gmail, Google Sheets, Airtable, HubSpot, Telegram.