715-line working bilingual SaaS prototype. EN and ES UI from day one. Drop-in governed front end for AI-as-a-service customers in LatAm and US-Spanish-speaking markets who need a customer-facing surface without spinning up their own — and who need governance, audit, and language reach in the same product.
WHL Thin Client gives AI-as-a-service operators a working customer-facing surface — login, conversation, document handling, audit export — without the multi-quarter build cost. Every inference is routed through the WHL governance gates. Every action is captured into a customer-visible audit record. The UI ships in English and Spanish from the first deployment.
English and Spanish surfaces from day one — not a translation backlog. Strings, prompts, audit-export labels, and customer-facing notices ship in both languages. Locale switch is per-user and persisted.
All model calls route through the WHL governance gates — proposal envelope, gate evaluation, dispatch, receipt emission. Refusals are explained in the user's locale. The audit chain is the source of truth, not the chat history.
Customer-controlled export of the audit record for a session, a user, or a date range. Hash-chained receipts, locale-aware labels, machine-readable schemas. Suitable for downstream regulatory submission or contractual evidence.
Per-customer dashboard showing usage, gate outcomes, overrides, anomaly state, and audit-export status. The customer sees their own surface — not the operator's internal console.
Tenant isolation, per-tenant configuration, per-tenant audit-chain partition. The 715-line prototype carries the wiring for multi-tenant operation; the operator chooses the deployment topology.
Built against the operational reality of Spanish-speaking enterprise and SMB deployments. The bilingual surface is a market-access feature, not an i18n afterthought.
The customer-facing surface accepts a brand skin — color, mark, locale defaults — so resellers can deliver a governed AI product under their own brand without rewriting the front end.
715 LOC running today. Trial deployments draw from the actual prototype. The roadmap from trial to production is incremental hardening — not a rewrite.
There is no "ungoverned mode." The inference path is the gated path. Operators choose policy strength per tenant; they do not choose whether the gate runs.
Thin Client is for the operator who has the model strategy, the customer pipeline, and the compliance commitment — and who does not want to spend two quarters writing a bilingual customer surface to start booking revenue.
Spanish-first surface, governed inference, audit export. Drop into Mexico, Colombia, Chile, Argentina deployments without translating a SaaS product designed for English-only markets.
Bilingual front end for US deployers serving Spanish-dominant customers in healthcare, financial services, and public-benefit programs. Locale per user. Audit record per language.
Brand the surface, keep the governed inference engine, ship in two languages. The reseller closes the customer; WHL Thin Client carries the front end and the audit chain.
WHL Thin Client is built for operators whose customer base spans Spanish-speaking markets and who treat audit and governance as a product feature.
Build verification of the bilingual thin-client prototype.
$ python -m build
* Building wheels for whl-thin-client
* Source: 715 LOC across 14 modules
* Locale support: EN, ES
* Hyperliquid client: verified
* WireGuard relay: verified
* Build: SUCCESSBuild verified: 715 LOC, bilingual EN/ES, working Hyperliquid client + WireGuard relay. Not yet hardened — read-only adjacent to live trading systems.
Trial deployments stand up the 715-LOC prototype against a representative tenant, exercise the bilingual surface, and walk through the audit-export workflow with your compliance team.