Source: 23-ENGINEERING-SCALE-QUALITY.md

What this source is

The full benchmark narrative behind the 60-second production-readiness snapshot on Start Here. Provides measured scale, measured quality, relatable executive benchmarks, and methodology. Verified April 5, 2026.

Key claims (new content beyond 04-technical-proof)

1. Relatable benchmarks for Squire executives (NEW framing):
   • Advisory/compliance complexity — comparable to Workiva, AuditBoard, Thomson Reuters/CCH platforms for evidence management, audit traceability, and governed approvals
   • iPaaS integration breadth — comparable to Celigo, Boomi, MuleSoft across connector behavior, orchestration, retries, monitoring, policy controls
   • Engineering rigor: “not prototype quality”
2. LOC by area (NEW breakdown of the ~854K total):
   • src/: 209,706 LOC
   • tests/: 148,168 LOC
   • docs/ + *.md: 226,285 LOC
   • public/: 126,091 LOC
   • Total: ~854K (matches Start Here)
3. Language footprint (NEW):
   • TypeScript snapshot: 362,598 LOC (total TS across repo — NOT the same as the 45,757 “production TypeScript” coverage baseline from 26-canonical-metrics-and-wording)
   • Markdown snapshot: 210,856 LOC
   • HTML snapshot: 182,416 LOC
4. Architecture signals, grep-based approximate (NEW — all concrete numbers):
   • ~1,035 route definitions
   • ~242 service files
   • ~378 DI (dependency injection) bindings
   • ~147 AI-related files
   • ~38 connector files
5. Test suite breakdown with suite counts (NEW detail): 384 unit suites, 20 integration suites, 2 E2E portal suites → 406 total suites. Prior sources had the total but not the breakdown.
6. Operational hardening highlights:
   • Hosted split architecture: Cloudflare Pages + Railway API (NEW — first formal mention of deployment topology)
   • Production guards for demo mode, CORS controls, rate limiting
   • Reasoning traces and compliance export paths
   • Dependency security remediation with residual-risk tracking
7. Methodology section — explicitly names the commands used: git ls-files, wc -l, npm run test:coverage:ci, npm run test:e2e, npm run test:e2e:portals. Approximate values marked with ~. This is strong transparency — a reviewer can reproduce the numbers.
8. Snapshot freshness note: tracked-file count refreshed 2026-03-06; LOC-by-area snapshot from March 2026 presentation build; test/coverage values verified April 5, 2026.

Pages updated by this ingest

Updated (2 existing pages):

• production-proof — added LOC-by-area breakdown, architecture signals (~1,035 routes / ~242 services / ~378 DI / ~147 AI / ~38 connectors), suite breakdown (384/20/2), hosted split deployment topology (Cloudflare Pages + Railway API), methodology references
• preston-test-repo — added the “comparable in complexity dimensions to…” executive benchmark framing

Cross-references / contradictions found

• “45,757 production TypeScript” vs “362,598 TypeScript total” — RECONCILED: no contradiction, just a subset. The 45,757 from 26-canonical-metrics-and-wording is the production TypeScript subset that has coverage measured against it (so 65% line coverage is measured against 45,757 lines). The 362,598 here is the TOTAL TypeScript across the whole repo (production + tests + infra + types + specs + generated). The 854K total text LOC includes TS + MD + HTML + everything else.
• “Comparable in complexity dimensions” wording is preferred per 26-canonical-metrics-and-wording style guide. This source uses exactly that phrasing — so the executive benchmark framing is style-guide-compliant by construction.
• Hosted split architecture (Cloudflare Pages + Railway API): first formally-ingested deployment topology fact. Earlier sources (04-technical-proof) mention Docker + Kubernetes + Terraform, which is the containerization/orchestration layer; this source adds the hosting providers.
• 384 unit suites + 20 integration suites + 2 E2E portal suites = 406 — matches canonical 406 total suites from 26-canonical-metrics-and-wording exactly. Full reconciliation.

Notes

• This source is designed to be read alongside 04-technical-proof — together they form the complete technical proof package. 04-TP gives features + AI providers + architecture; 23-ESQ gives scale + quality + benchmarks + methodology.
• The grep-based “approximate” values (~1,035 routes etc.) are honest about their measurement technique. A reviewer could run the same greps and verify.
• 1,035 route definitions is a significant number for a 406-suite test base — roughly 3.9 routes per test suite on average, which is consistent with meaningful integration testing coverage.
• 147 AI-related files and 38 connector files — the AI surface is ~4× larger than the connector surface in file count, consistent with SuiteCentral 2.0 positioning itself as an AI-first platform with connectors as a substrate.