İçeriğe geç
CodEven/Demos/n8n Library
Production Automation Recipes

n8n Workflow Library

5 importable, production-tested n8n workflows featuring schema validation, resilient error branches, and in-browser execution simulations.

AI & CRM7 nodes
OpenAIHubSpotSlackWebhook

AI Lead Qualification -> HubSpot & Slack

Inbound form webhook parsed with LLM scoring, JSON schema validation, automatic contact upsert, and priority Slack alerts for hot leads.

Trigger:Webhook (POST /lead-qualification)

Visual Node Pipeline

Webhook Trigger
Webhook - Inbound Lead
LLM Scoring
OpenAI - Qualify & Score Lead
Schema Validation
Validate Schema & Classify
Contract Gate
Is Schema Valid?
HubSpot CRM
HubSpot - Upsert Contact
Priority Check
Is Hot Lead?
Slack Notification
Slack - Alert Sales Team
Error Handling & Fallbacks

If the LLM returns invalid JSON or unexpected fields, the validation node intercepts the error, flags the record with 'needs_review', and assigns it to a manual review queue in HubSpot instead of dropping the submission.

Audit & Credentials

Credential references are abstracted with placeholder names. No hardcoded secrets, ready to bind with your workspace credentials.

Download JSON
Import directly into n8n via Settings → Workflows → Import
Finance & OCR6 nodes
IMAP / EmailOpenAIGoogle SheetsSlack

Invoice OCR -> Google Sheets with Math Audit

Email attachment trigger extracts PDF text, structures financial figures via LLM, validates subtotal + VAT = total math, and appends rows to Google Sheets.

Trigger:IMAP Email Trigger (Attachment PDF)

Visual Node Pipeline

Email Trigger
IMAP Email - Inbound Invoices
PDF Parser
Extract PDF Text
Field Extraction
OpenAI - Extract Financial Fields
Math Audit Gate
Validate Total vs VAT Math
Validation Gate
Is Math Valid?
Google Sheets
Google Sheets - Append Invoice Row
Error Handling & Fallbacks

A strict math check validates that subtotal + VAT matches the stated total within a 0.05 rounding tolerance. Any mismatch redirects the entry to a Slack human approval queue.

Audit & Credentials

Credential references are abstracted with placeholder names. No hardcoded secrets, ready to bind with your workspace credentials.

Download JSON
Import directly into n8n via Settings → Workflows → Import
Retention & Reviews7 nodes
PostgreSQLWhatsApp Cloud APISlackWebhook

Post-Appointment Review & Sentiment Routing

Hourly cron queries appointments completed 2 hours prior, sends interactive WhatsApp survey, and routes 4-5 star ratings to Google Maps and 1-3 star ratings to managers.

Trigger:Schedule Trigger (Hourly Cron)

Visual Node Pipeline

Hourly Cron
Schedule Trigger - Hourly
Query Completed
Postgres - Fetch Completed Appointments
Send Survey
WhatsApp - Dispatch 2h Followup Survey
Rating Inbound
Webhook - Inbound Customer Rating
Parse Rating
Parse Rating & Sentiment
Sentiment Gate
Is High Rating (4-5 Stars)?
Google Review Link
WhatsApp - Send Google Review Link
Error Handling & Fallbacks

Dissatisfied customers (1-3 stars) never receive public review links. Instead, their feedback triggers an escalation alert directly to the manager's channel for same-day resolution.

Audit & Credentials

Credential references are abstracted with placeholder names. No hardcoded secrets, ready to bind with your workspace credentials.

Download JSON
Import directly into n8n via Settings → Workflows → Import
Operations6 nodes
HTTP / APIOpenAISMTP / Email

Daily Operations Briefing -> Executive Email

Cron at 08:00 fetches yesterday's lead, appointment, and support metrics in parallel, synthesizes a 5-bullet executive summary with LLM, and sends an HTML briefing email.

Trigger:Schedule Trigger (Cron 08:00 AM)

Visual Node Pipeline

Morning Cron
Schedule Trigger - 08:00 AM
Lead Metrics
HTTP - Fetch Yesterday Leads
Aggregate Metrics
Aggregate Daily Metrics
LLM Briefing
OpenAI - Synthesize 5-Bullet Summary
HTML Layout
Format HTML Email Body
Send Email
Send Briefing Email
Error Handling & Fallbacks

If individual metrics endpoints timeout or return partial payloads, fallback defaults preserve the morning briefing run so founders always receive an updated operational snapshot.

Audit & Credentials

Credential references are abstracted with placeholder names. No hardcoded secrets, ready to bind with your workspace credentials.

Download JSON
Import directly into n8n via Settings → Workflows → Import
Infrastructure8 nodes
PostgreSQLWhatsApp Cloud APIOpenAIn8n Wait Node

Messaging Burst Debounce & Context Guard

Prevents concurrent duplicate LLM replies when users send rapid consecutive messages. Uses a 4-second wait, Postgres count comparison, and explicit source node referencing.

Trigger:Webhook (Inbound WhatsApp Message)

Visual Node Pipeline

Webhook Inbound
Webhook - Inbound WhatsApp Message
Verify Secret
VerifySecret
Persist Inbound
SaveInboundMessage
Wait 4 Seconds
DebounceWait
Check Newer Messages
CheckNewer
Is Latest Execution?
IsNewest
Fetch Conversation History
FetchHistory
Dispatch Unified Reply
WhatsApp - Send Unified Reply
Error Handling & Fallbacks

If the CheckNewer query cannot locate the triggering record, the subquery returns NULL and newer defaults to '0', allowing the bot to reply rather than staying silent.

Production Lesson from Real Incident

Real production lesson: When intermediate nodes were inserted into the flow, FetchHistory's query originally read conversationId from implicit $json (which now held CheckNewer's { newer: '0' } output). This caused the history query to silently return 0 rows while reporting success. The fix was anchoring the expression explicitly to $('Webhook - Inbound WhatsApp Message').item.json.

Download JSON
Import directly into n8n via Settings → Workflows → Import

Need production-ready n8n automations in your stack?

Let's build resilient data pipelines with strict schema contracts, defensive error handling, and clean credential separation.

    n8n Workflow Library — Production Automations & Integrations | CodEven