Back to Resources
Resources

The AI Tool Stack We Use to Run 8 Companies

Our complete technology stack across 8 ventures. Workflows, databases, AI models, hosting, and voice — with real costs and honest alternatives.

8 min read
tool stackAI toolsn8nSupabaseOpenRouter

We run eight ventures — CDL Schools USA, TaxLienSimple, SceneHost, VettyDrive, SchoolRegistry.ng, NigeriaPolls, AckPost, and BluprintCreations itself. Each has different needs, different budgets, and different technical complexity.

We don't use one stack because we sell one. We use this stack because it works across high-volume content sites, regulated training platforms, fleet management software, civic data aggregation, and social publishing — all at the same time.

Below is our complete stack. For each tool: what we use it for, why we chose it, what it costs us, and what we'd use instead if we had to switch tomorrow.


Workflows: n8n, Make.com, Zapier

What we use them for:

  • n8n (self-hosted): Heavy data processing, complex conditional logic, systems handling 10,000+ executions monthly. Powers SchoolRegistry's content pipeline, NigeriaPolls data aggregation, and VettyDrive's maintenance scheduling.
  • Make.com: Visual workflow building for marketing ops — content distribution, lead nurturing sequences, CRM updates. Used by SceneHost and TaxLienSimple.
  • Zapier: Quick connections to niche apps, one-off integrations, and client handoffs where simplicity matters. AckPost uses it for certain social platform APIs.

Why we chose them: n8n gives us control and cost efficiency at scale. Make.com hits the sweet spot between power and accessibility. Zapier has the largest app directory on earth. We don't pick favorites — we pick the right tool for the workflow.

What it costs: n8n self-hosted ~$12/month VPS per instance. Make.com $9-16/month per team. Zapier $19.99-49/month per account (we minimize usage).

Alternatives we'd consider: Pipedream (strong code-first alternative to n8n), Workato (enterprise), ActivePieces (open-source, growing fast).


Database: Supabase, PostgreSQL

What we use them for: Supabase powers every venture's data layer — user accounts, content storage, analytics, and real-time features. VettyDrive's fleet records, SchoolRegistry's school profiles, SceneHost's property data, CDL Schools' lead pipeline — all live on Supabase.

Why we chose it: PostgreSQL is battle-tested. Supabase adds auth, real-time subscriptions, storage, and edge functions without the Firebase lock-in. It's open-source, self-hostable if needed, and pricing is transparent. We can export our data anytime.

What it costs: Free tier for development. Pro: $25/month per project. Some ventures run multiple projects; total database spend across all 8: ~$200/month.

Alternatives we'd consider: Direct PostgreSQL on VPS (more control, more maintenance), Neon (serverless PostgreSQL), PlanetScale (MySQL-compatible, excellent for content-heavy sites).


AI Models: OpenRouter, Claude, GPT-4, Gemini

What we use them for:

  • OpenRouter: Universal API gateway. One endpoint, access to 50+ models. We use it for routing traffic to the best model for each job without managing multiple API keys.
  • Claude 3.5 Sonnet: Long-form content generation, document analysis, complex reasoning. Our go-to for TaxLienSimple articles and SchoolRegistry page content.
  • GPT-4o: Fast, general-purpose tasks — summarization, classification, light content generation. Used across all ventures for quick wins.
  • Gemini 1.5 Pro: Massive context window (1M+ tokens) for processing long documents, legal text, and large codebases. Used for CDL Schools' compliance document analysis.

Why we chose them: Different models excel at different tasks. Claude writes better long-form. GPT-4o is fastest for simple tasks. Gemini handles context no other model can touch. OpenRouter lets us switch between them without code changes.

What it costs: Highly variable by usage. TaxLienSimple (heaviest user): ~$400/month in AI costs for 50+ articles. Most other ventures: $50-150/month. Total AI spend across 8 ventures: ~$1,200/month.

Alternatives we'd consider: Direct Anthropic/OpenAI APIs (if we didn't need model flexibility), Groq (for ultra-low-latency), Local models via Ollama (for sensitive data, cost reduction).


Content: Custom CMS, SEO Tools

What we use them for: Our content engine is built, not bought. Custom CMS handles TaxLienSimple's 50 articles/month, SchoolRegistry's 15,000 pages, and SceneHost's property guides. SEO tooling is integrated directly — automated meta generation, internal linking, schema markup, and rank tracking.

Why we chose this approach: No off-the-shelf CMS handles 15,000 auto-generated pages well. WordPress chokes. Contentful gets expensive. Our stack is purpose-built for AI-generated content at scale with clean HTML, fast rendering, and programmatic SEO baked in.

What it costs: Custom CMS development cost amortized; hosting negligible (~$20/month). SEO tooling: Screaming Frog (one-time), Ahrefs ($99/month), custom rank trackers (~$50/month). Total: ~$170/month.

Alternatives we'd consider: Astro + Markdown (for simpler content sites), Sanity (structured content, good for multi-channel), Strapi (open-source, self-hosted CMS).


Phone & Voice: Bland.ai, Vapi.ai

What we use them for: Bland.ai for AI phone calls — lead qualification, appointment setting, follow-up calls. Sounds human, handles objections, books directly into calendars. Vapi.ai for voice AI — reservation confirmations, maintenance reminders, guest check-in calls.

Why we chose them: Voice AI is the highest-ROI channel most businesses ignore. Bland.ai has the most natural-sounding voices we've tested. Vapi.ai offers more customization for complex call flows. Both are API-first and integrate cleanly with our workflow stack.

What it costs: Bland.ai ~$0.09/minute. CDL Schools runs ~2,000 minutes/month = $180. Vapi.ai ~$0.05/minute. Lighter usage across ventures = ~$80/month. Total voice AI: ~$260/month.

Alternatives we'd consider: Retell.ai (strong competitor to Vapi), Synthflow (good for simpler use cases), Twilio + custom Whisper (if we needed full control).


Hosting: VPS, Cloudflare

What we use them for: VPS (Hetzner, DigitalOcean) for primary hosting — n8n instances, databases, APIs, and custom applications. We self-host where it makes sense and use managed services where it doesn't. Cloudflare for CDN, DDoS protection, edge caching, DNS, and edge functions. Every venture runs behind Cloudflare.

Why we chose them: Hetzner offers the best price-to-performance ratio in Europe. DigitalOcean is reliable for US-based projects. Cloudflare is non-negotiable for speed and security at our scale.

What it costs: Hetzner €5-20/month per server (we run 4). DigitalOcean $12-48/month per droplet (we run 3). Cloudflare Free tier mostly; Pro on 2 ventures ($20/month each). Total hosting: ~$250/month for 8 ventures.

Alternatives we'd consider: AWS/GCP (if a client demanded it), Vercel (for frontend-only projects), Railway/Render (for simpler deployment, higher cost).


Total Stack Cost (All 8 Ventures)

CategoryMonthly Cost
Workflows (n8n/Make/Zapier)~$150
Database (Supabase)~$200
AI Models~$1,200
Content/CMS/SEO~$170
Phone/Voice AI~$260
Hosting/VPS~$250
Total~$2,230/month

Per-venture average: ~$279/month. Compare that to the cost of even a single full-time developer, content team, or social media agency. The leverage is absurd — but only if the stack is built right.


Related Resources

BluprintCreations — We eat our own cooking. Every tool listed here runs our own ventures. No affiliate links. No kickbacks. Just what works.

Get This Stack Built For Your Business — Book a Fit Call

Book a Fit Call