Our stack

We preach owning the machine. Here’s ours.

StackSwap replaces rented GTM sprawl with one engine you own: a buyer model trained on your data, scoring logic you can inspect, and outcomes that compound instead of expiring at renewal. This page is the receipt — the product surfaces already shipped, the data layer underneath, and the rented rails we use to move fast without giving up control.

Owned where it matters. Rented where it should be.

What this page proves

Most GTM vendors ask you to trust a black box: data goes in, a number comes out, and the scoring logic stays locked in someone else’s account. StackSwap’s bet is the opposite. The model, the scoring logic, the data flows, and the recommendations should be inspectable, portable, and explainable. So we built ours that way — and this page documents it in the open.

Real

Built on a real production stack, not a mockup

Every card below links to a long-form write-up of something that exists in the codebase today — runtimes, schemas, and product surfaces, not slideware.

Live

The recommendation engine and knowledge base are already shipped

The free tools, comparison pages, and StackScan flow run in production now. The StackSwap OS product itself is still pre-beta (closed sandbox beta opens July 1, 2026).

Powering

The same data model already powers audits, reports, and comparisons

One canonical layer feeds the StackScan audit, the generated GTM report, and the premium comparison pages — so the numbers stay consistent across every surface.

Compounds

The machine behind the free tools becomes the intelligence layer in StackSwap OS

The engine that scores and explains public stacks is the same engine that becomes your owned buyer model inside StackSwap OS — it compounds instead of resetting at renewal.

Owned where it matters: data model, scoring logic, recommendation engine, product surfaces, and deployment code.
Rented where it should be: payments, hosting, LLM providers, and infrastructure rails.

The machine we own

Product surfaces already live — not roadmap claims.

🧩

StackScan Flow

URL-first intake, Supabase-backed lead capture, and unified GTM Engine results route.

Read →
🚀

Recommendation Engine

7 stack blueprints, 6 fit rules, company-specific stack recommendations via Claude.

Read →
🎯

Tool Capability Map

16 tools with functions, strengths, limitations, replacement rules, fit logic.

Read →
📈

AI-Native Gap Metric

Measures legacy exposure as a percentage. Color-coded. Benchmarked against industry.

Read →
📄

GTM Report Template

10-page dark-mode GTM strategy HTML template and TypeScript generator wired for PDF export.

Read →
📊

Premium Comparison Pages

13 /compare/[slug] pages with Article + FAQPage JSON-LD, structured verdicts, and editorial tuned for LLM-SEO.

Read →
📚

Knowledge Base Engine

Comparison hub, category pages, 60+ articles. FAQ schema + canonical URLs drive LLM-SEO across the entire KB.

Read →
🔗

Cross-Tool Data Flow

StackScan, StackBuilder, StackEnrich, and GTM Strategy pass context via URL params + sessionStorage.

Read →
🌐

Company URL Inference

Server-side company lookup API that infers name, industry, positioning, and competitors from a single URL.

Read →
📋

Sortable Plan Table

Tri-state column sort with affordance arrows. Shared component between paid, free preview, and pre-paywall surfaces.

Read →
🧬

Operating Profile Inference

Unified profile object for demo + real users, powering tailored executive summary and benchmark credibility lines.

Read →
🧱

StackScan Data Adapter

Adapter that maps StackScan / StackBuild analysis and form context into a single GTM strategy report model.

Read →
💳

Report Checkout + Unlock Flow

Stripe Checkout for one-time $99 report unlocks, webhook verification, and report-level premium state.

Read →

The data layer underneath

The systems that make recommendations consistent, scored, and auditable.

🗂

Supabase Data Layer

21-column schema. RLS policies. Source tagging. Organic intelligence views.

The shared store behind every scan, report, and benchmark, so recommendations stay consistent across surfaces.

Read →
💰

1,120-Tool Database

Per-seat + flat-rate pricing. Team multipliers. Industry benchmarks. 837 priced.

The priced catalog behind overlap detection, swap suggestions, and cost math.

Read →
⚖️

11-Source Weighted Pricing Council

11 independent SaaS pricing datasets merged with confidence weighting behind every tool cost estimate.

Combines multiple pricing sources with confidence weighting, so estimates aren’t based on one stale number.

Read →
💵

Canonical Money Architecture

One-number-across-surfaces rule: hero, plan table, stack transformation, and CTA all read the same canonical source.

One source of truth for pricing math, so different pages can’t show different cost estimates.

Read →
📬

Leads Schema v2

Extended Supabase leads table for StackScan: GTM fields, analysis scores, spend estimates, source tagging.

One structured record per submission, so analytics and reporting stay clean and segmentable.

Read →
📏

Team-Size Precision Pipeline

Precise numeric headcount threads through scanStack end-to-end. No more lossy band round-trip between intake and scan.

Carries your exact headcount through the whole scan, so a 15-person team and a 45-person team get different math.

Read →
🔄

Circular Redundancy Safeguards

Dedup logic blocks "remove A because B / remove B because A" rationales from ever reaching the execution timeline.

Prevents the engine from recommending circular or duplicate replacement logic.

Read →
🧪

Recommendation Normalization Layer

Deterministic validation pass that resolves conflicting KEEP/REPLACE/REMOVE/ADD outcomes before render.

Resolves conflicting keep/replace/remove calls before you see them, so the plan stays internally consistent.

Read →
🎲

StackScan Fuzz Harness

500-scan invariant sweeps across 12 rules. Catches pricing drift, circular logic, and phantom savings before users do.

Runs hundreds of synthetic stacks through the engine before launch, so the savings math is proven consistent before you see it.

Read →
📊

Internal BI Dashboard

Real vs seed filtering. Funnel analysis. Password-gated.

Separates real traffic from seed data, so internal metrics and public claims stay honest.

Read →

Platform.js

Shared nav + footer, metrics, Supabase, count-up animations. Pufferfish.

Shared navigation, footer, and data wiring, so every surface behaves like one product, not a pile of apps.

Read →
🎨

main.css Design System

Single CSS source of truth. Cards, buttons, inputs, metrics, pills, grids. All 14 pages.

A single styling source of truth, so every page stays visually consistent as new ones ship.

Read →

The rails we rent

Replaceable infrastructure we rent to stay fast, portable, and cheap to run.

“The operators who learn what makes their buyers convert — and compound that knowledge every month — are going to eat.”