Workflow Automation & Marketing OperationsPROJECT 12 — SUPPORTING SYSTEMVerified Repository & SQLite FTS5 Index
Growth Flow HubSelf-Hosted n8n Workflow Collection & Searchable Automation Catalog
A production-grade, self-hosted automation ecosystem turning a vast library of complex n8n workflow JSON pipelines into an instantly searchable, categorized, and exportable engineering catalog powered by FastAPI, SQLite FTS5, and Docker.
The Business Problem
Modern marketing and growth operations rely on dozens of disparate SaaS applications—ad networks, CRM databases, email deliverability tools, and LLM APIs. Connecting these through custom point-to-point glue code creates high maintenance overhead, fragile single points of failure, and zero central visibility.
Conversely, when teams adopt low-code workflow platforms like n8n, they rapidly accumulate hundreds of raw JSON workflow files. Without a structured catalog or discovery layer, developers and growth engineers face severe friction:
Zero Searchability
Raw JSON export files cannot be searched across node configurations, webhook endpoints, or credentials without opening every file manually.
Duplication & Drift
Different growth pods recreate the exact same Meta CAPI or lead-enrichment workflows from scratch rather than reusing hardened templates.
Security Vulnerabilities
Unvetted community workflows frequently include unvalidated webhook params, unencrypted API tokens, and missing CORS or rate limits.
Execution Opacity
When webhook payloads fail silently, growth teams lose critical conversion events and attribution tracking without immediate alerts.
What I Built: Growth Flow Hub
I engineered a production catalog and API microservice that transforms thousands of isolated n8n workflow JSON files into an indexed, discoverable product surface. Backed by FastAPI, an SQLite FTS5 (Full-Text Search) engine, and Docker containerization, it enables instantaneous keyword lookup, complexity filtering, and verified JSON exports.
FTS5 Search Index
Tokenizes all node types, descriptions, triggers, and integration tags into an embedded SQLite full-text search table for sub-50ms queries.
FastAPI REST Layer
Clean REST endpoints (/api/search, /api/workflow/:id, /api/export/:id) with Pydantic schema validation and rate limiting.
Hardened Delivery
Strict path traversal sanitization, CORS allowlisting, non-root Docker execution user, and automated Trivy vulnerability checks.
How It Works: End-to-End Pipeline
Ingestion → Tokenization → Query Execution → Secure JSON Delivery
01
Raw JSON Collection & Normalization
Pipelines created in n8n (or authored as JSON specifications) are ingested from the static repository structure. Each file is parsed to extract metadata: node types, trigger mechanisms, and integration connectors.
02
SQLite FTS5 Metadata Indexing
A standalone Python indexing pipeline (`workflow_db.py`) creates a high-performance SQLite virtual table (`workflows_fts`). Text fields are tokenized with Porter stemming for rapid substring and semantic matching.
03
Multi-Dimensional Filter Narrowing
The FastAPI query parser evaluates parameters: `category` (Marketing, SEO, DevOps), `trigger` (Webhook, Cron, Poll), `complexity` (Beginner to Expert), and `services` (Stripe, Meta, Slack, OpenAI).
04
Sanitized Inspection & Direct n8n Import
Engineers view full node topologies, credential requirements, and parameter mappings in the browser UI, then copy or export clean JSON directly into their local or cloud n8n instances.
Verified Production Artifacts
High-Resolution Screenshot of Active n8n Automation Mesh
Full Resolution
Production n8n Growth Pipeline Canvas
Real-world n8n execution canvas illustrating webhook data ingestion, conditional data branching, AI enrichment, and multi-endpoint notifications.
Interactive Workflow Explorer
Sample Workflows from the Catalog Repository
WF-01Webhook (POST)
Meta CAPI Server-Side Lead Ingestion & Deduplication
Receives form webhooks, hashes PII (SHA-256), checks SQLite dedup table, and dispatches to Meta Conversions API.
Meta Graph APIPostgreSQLSlackCrypto Hash
14 NodesAdvanced
WF-02Cron (Daily 06:00 UTC)
SERP Rank Drop Trigger → Claude Content Refresh Brief
Polls Google Search Console performance data for >3 position drops, queries Anthropic for remediation brief, and notifies editorial desk.
GSC APIClaude 3.5 SonnetNotionSendGrid
18 NodesAdvanced
WF-03Stripe Webhook
CRM Multi-Touch Attribution Event Dispatcher
Intercepts charge.succeeded events, correlates anonymous session cookie ID, and attributes touchpoint weights.
StripePathfinder MTA APIHubSpotPostgres
11 NodesIntermediate
WF-04FastAPI REST Endpoint
Autonomous Multi-Agent Task Routing Mesh
Parses complex natural-language business requests, assigns work to departmental agent pods, and aggregates responses.
FastAPIOllama / Local LLMAgents OfficeGitHub
22 NodesExpert
WF-05Webhook (Typeform)
High-Volume Lead Enrichment & Domain Validation
Validates corporate email MX records, checks company employee count, and tags SDR priorities in real-time.
Clearbit APIHunter.ioAirtableSlack
9 NodesIntermediate
WF-06Interval (Every 5 mins)
Automated Playwright Health & Uptime Ping Probe
Executes simulated headless browser login flow; triggers PagerDuty escalation if selector fails within 5,000ms.
Playwright HeadlessPagerDutyWebhookDiscord
8 NodesIntermediate
Technical Architecture & Source Verification
Verified from api_server.py, workflow_db.py & Dockerfile
The catalog microservice separates data storage, search querying, and frontend rendering into cleanly decoupled layers:
// Modular Architecture Tree
├── api_server.py → FastAPI application routing, CORS, rate limiting, Pydantic schemas
├── workflow_db.py → SQLite FTS5 database wrapper, tokenizer, and query builder
├── workflows/ → Categorized directory storing raw, validated JSON pipeline files
├── static/ → Responsive single-page application (HTML, CSS, JS) for instant search
└── Dockerfile → Multi-stage production container running as non-root user (UID 10001)
Verified Claim Boundaries & Evidence Limits
Ground-Truth Engineering Assertions (Source Seed Compliance)
Workflow Count Documentation: The source repository README reports 4,343 workflow templates across 365 integrations and 29,445 nodes (an earlier snapshot referenced 2,060+). These figures are versioned catalog metrics and do not claim that every pipeline has been executed in live client production.
Operational Catalog vs. Campaign Performance: This system is an operational workflow library and search platform. It does NOT claim automated lead volume, conversion lift, or revenue generated from workflows alone.
Credential & Environment Independence: Importing any workflow requires users to configure their own API credentials (OAuth tokens, webhook endpoints, database URIs). The catalog provides verified schemas and node topologies.
Ecosystem & Cross-Project Integrations
How Growth Flow Operationalizes Other Growth Systems
Project 03 — GSC Cockpit
Search Console Performance Dashboard
Supplies rank-drop and CTR-anomaly triggers that initiate n8n content triage webhooks.
Project 04 — OmniGrowth
AI Marketing Operations Platform
Executes marketing workflow pipelines deployed from Growth Flow templates across CRM and ad channels.
Project 07 — Marketing MCP
Marketing API Automation MCP
Exposes Growth Flow search and workflow retrieval endpoints as MCP tool calls for Claude and Cursor.
Project 10 — Agents Office
Local Multi-Agent Workstation
Virtual office agents trigger pre-tested n8n background automations for recurring business workflows.
Recruiter Briefing5-Second Scan
Role RelevanceMarketing Automation Engineer, Growth Operations, Systems Integrator
Core CapabilitySelf-hosted workflow cataloging, SQLite FTS5 search engine, Python/FastAPI microservices, Docker container security.
Verified Tech Stack
Python 3.11FastAPISQLite FTS5Dockern8nPydanticWebhooksREST API
Verified Artifacts
Source CodeVerified GitHub
FTS5 SearchTested (<50ms)
Security HardeningNon-Root Docker
The Growth Lifecycle Mesh
Growth Flow functions as the connective tissue across the portfolio: feeding alerts from Search Intelligence into Content Generation, and piping attribution data back to CRM databases.
Available for 2026 Roles
Let's build repeatable growth engines.
I'm open to Digital Growth, Technical SEO, SaaS, AI Automation, and Analytics opportunities.