101 citations named eleven HIPs that no longer exist. They were removed deliberately — 3a52264d gives a reason for each, either that shipping code contradicts the document or that the thing was never built — so the references wanted repointing at whatever survived, not the files back. Each survivor was established from content and confirmed against the candidate's own title and text, never guessed from the number. Object storage went to HIP-0405, whose CR carries the same repository, endpoint and port pair the deleted standard specified. Observability went to HIP-0132, which corrects the deleted claim by name: one telemetry plane per node, not a sidecar. Where nothing survived because nothing was built, the sentence was reworded to stand on its own rather than given an invented citation. Two citations turned out not to be about their stated subject at all. HIP-0033 cited "HIP-0032 (CI/CD Standard)" — a number that was never CI/CD; the subject is HIP-0036, which already appeared as the child of the same diagram node. HIP-0107 cited it as "the vfs.Backend interface contract this HIP standardizes on", but the deleted text contains no mention of vfs and HIP-0107 defines that interface itself. (cherry picked from commit 32989a78d640e61446e26151892f1d8c4a3dd4b3)
4.8 KiB
hip, title, author, type, category, status, created, requires
| hip | title | author | type | category | status | created | requires |
|---|---|---|---|---|---|---|---|
| 0115 | Hanzo Frontend Delivery | Hanzo AI Team | Standards Track | Infrastructure | Draft | 2026-06-25 | HIP-0036, HIP-0068, HIP-0119, HIP-0504 |
HIP-0115: Hanzo Frontend Delivery
Abstract
This is the one and only way a frontend — a site, a docs surface, a dashboard — is organized, built, and shipped across Hanzo, Lux, and Zoo. Every frontend is one repo in its org's apps organization, is one application on the PaaS, builds the same way, and is served the same way. There is exactly one project per site; there are no second implementations and no second homes.
It is the frontend peer of HIP-0119. HIP-0119 governs backend services — they
serve /v1/* and :9090 to machines, live in per-service repos, and are deployed
by the operator. This HIP governs apps — they serve / to humans, live in the
per-org apps organization, and are deployed by the PaaS. The split is the
decomplection: machine surface vs human surface, operator vs PaaS, never braided.
Specification
The key words MUST, MUST NOT, SHOULD, SHOULD NOT, MAY are per RFC 2119.
§1 One repo per site, in the org's apps organization
- Every frontend is its own repository in the
<org>-appsGitHub organization (hanzo-apps,lux-apps,zoo-apps), cloned locally to~/work/<org>/apps/<name>. - One project per site. A site MUST have exactly one repository. Multiple
implementations of the same site (e.g.
docs,app-docs,platform-docs) MUST be collapsed to one; the others are retired (archived), not left to rot. - A frontend MUST NOT live in the platform/service org (
hanzoai,luxfi,zooai). Those orgs hold backend services and libraries. Static content and sites that currently sit there MUST be migrated into the apps org. One home. - Repos are separate (not a monorepo) precisely so each app is an independent PaaS application with an independent deploy. Sharing is by published package (the design system), never by workspace coupling.
§2 The design system is shared, not copied
A frontend MUST consume the design system as a dependency — @hanzo/ui, themed
per brand to @luxfi/ui / @zooai/ui (HIP-0504: brand UIs are theme values, not
forks). A frontend MUST NOT vendor or re-implement shared components. Brand differs
in values (tokens, logo, domain); shape is identical.
§3 Build — the PaaS, one of two targets
Every app builds on the PaaS (platform.hanzo.ai). An app declares exactly one
build target:
static— the default for sites and docs. The PaaS builds the app and publishes the static export (Nextoutput: export, Vitedist, etc.) as an immutable artifact. This is the common case; prefer it.container— only when the app genuinely needs a server at request time (SSR, server actions, auth callbacks that can't be edge/static). Runs as a PaaS container app.
An app MUST NOT be built or deployed by a hand-run pipeline, a bespoke per-repo GitHub Action that pushes to a host, or a second PaaS. One control plane.
§4 Serve — static to s3 + ingress, no nginx
- A
staticapp's artifact MUST be published tohanzoai/s3(the object store) and served byhanzoai/ingress+ the static plugin. nginx, caddy, and per-site web servers MUST NOT be used (per the platform rule). - A
containerapp is served throughhanzoai/ingress→ the app, per HIP-0068. - DNS is Cloudflare; TLS is cert-manager/ingress. The host is
<site>.<brand-domain>. - Adding a site is: new repo in
<org>-apps→ register the PaaS app (static, host) → deploy. No cluster YAML hand-editing, no new web server.
§5 One uniform shape across Hanzo, Lux, Zoo
hanzo-apps, lux-apps, and zoo-apps are the same organization in three brands.
An app in any of them follows this HIP identically; only values differ — the org
segment (<org>-apps), the brand domain, the theme package. LPs and ZIPs adopt
this HIP by reference and MUST NOT restate or fork it.
§6 Forbidden
- A site implemented in more than one repo; a site living in the service org.
- A frontend that vendors the design system instead of depending on it.
- A bespoke deploy pipeline; a second PaaS; nginx/caddy/per-site servers.
- Static artifacts served from anywhere but
hanzoai/s3via ingress+static.
Conformance checklist
- The site is exactly one repo in
<org>-apps; no alternate implementations. - It depends on the design system; it vendors nothing shared.
- It builds on the PaaS as
static(orcontaineronly if it must). staticartifacts go tohanzoai/s3, served by ingress+static — no nginx.- One host, Cloudflare DNS, ingress TLS.
- Nothing from §6.
References
- HIP-0036 CI/CD Build System · HIP-0068 Ingress Standard
- HIP-0119 Service Conventions (the backend peer) · HIP-0504 Design System