Author SEO titles for the 42 English pages whose title tag is a single word

[hotlong]

Blocked-by: #171

Why

#166 added the mechanism — an optional seoTitle frontmatter field that feeds the title tag while the H1 and sidebar keep the short noun. It merged as 23a976b. This card fills it in.

The field is named seoTitle. #166 specified the mechanism but not the name; the implementing dev chose it and it is now the contract. It is schema-tightened with .trim().min(1), so an empty or whitespace-only value fails the build rather than silently producing a bare | ObjectOS.

42 of 79 English pages currently emit a one-word title tag (Views, Actions, Pages, Forms, Agents, Storage, Email, Permissions, Positions, Records, …). Rendered through the %s | ObjectOS template these match no query anyone types. Nobody searches "views"; people search "kanban view low code", "salesforce-style list views self hosted", "internal tool record layout".

The raw material already exists

Every one of those 42 pages carries a written, specific description. Two examples from main:

• content/docs/build/interface/views.mdx — title Views, description "List, Form, Kanban, Calendar, Gantt and more — how every object surface is declared."
• content/docs/build/data/relationships.mdx — title Relationships, description "Connect objects with lookups and master-detail — cascade rules, filtered pickers, hierarchies, junction objects, and roll-ups."

The terms a searcher would use are in the description. The task is largely to lift them into the title field, not to invent new copy. Read each page's description and body before writing its title — a title that promises something the page does not deliver costs more than a weak title, because it converts into a bounce.

What to change

• For each English page whose effective title tag is two words or fewer, declare a seoTitle. Keep it under roughly 60 characters so it is not truncated in the result page, and account for the | ObjectOS suffix the template appends.
• Lead with the specific term, not the brand — the template already supplies the brand.
• Do not change the title field itself. The sidebar, breadcrumbs and H1 read better as short nouns and that is why generateMetadata returns four fields: docs pages ship no SEO title, no Open Graph, and no structured data #166 separated the two.
• Propagate to the six translated locales only for pages that already have a translation file for that locale. Do not create new translation files here. Translated SEO titles must be written in the target language against local search vocabulary, not transliterated from English.

Verification

• After the change, no English page emits a title tag of two words or fewer except where a one-word title is genuinely the strongest choice — list any such page and its reason in the PR body.
• Every locale file touched still parses and its title tag renders in its own language.
• No content/docs/** file gains or loses a title, description, or body line.

---

Card corrections, made when the blocker was re-pointed:

• The blocker moved from generateMetadata returns four fields: docs pages ship no SEO title, no Open Graph, and no structured data #166 (merged) to [Decision] What does docs.objectos.ai target — the root URL is a redirect, the index title is "ObjectOS | ObjectOS", and the brand is spelled three ways #171. See the comment below for why.
• An earlier version of this body said the missing-translation gap "is its own card, Sitemap and hreflang advertise seven locales for pages that exist only in English — 41 of 79 pages #169". That was a wrong reference — Sitemap and hreflang advertise seven locales for pages that exist only in English — 41 of 79 pages #169 was the sitemap/hreflang advertisement card and has merged. The translation gap itself has no card; Docs ships seven locales and the marketing site ships eight — zh-Hant is the missing one, and docs already has a comment waiting for it #172 covers adding zh-Hant only.

[hotlong]

Blocker re-pointed: #166 → #171. Still pm:blocked, but by a different card, and it is worth being explicit that this is not a formality.

#166 merged as 23a976b, so the mechanism this card needs exists. What is missing now is not code — it is the answer to what these 42 titles should target, and that is #171's open question: does docs.objectos.ai compete for head terms, or cede them to www.objectos.ai and aim only at documentation-intent long tail?

The two answers produce genuinely different titles for the same page. For content/docs/build/interface/views.mdx:

• Ceding head terms (my recommendation on [Decision] What does docs.objectos.ai target — the root URL is a redirect, the index title is "ObjectOS | ObjectOS", and the brand is spelled three ways #171) — target someone who already knows the product and wants the how: Views: List, Kanban, Calendar & Gantt — ObjectOS.
• Competing for head terms — target someone shopping for a category: Kanban & List Views for Self-Hosted Internal Tools.

Writing 42 of these the wrong way and rewriting them later is worse than waiting, because titles accrue whatever ranking they earn and churning them discards it.

Not everything here is blocked, but splitting the card is worse than waiting. Some pages are unambiguous either way — reference/environment-variables.mdx, reference/cli.mdx and most of operate/** are documentation-intent under any strategy. But a partial pass would leave the file half-annotated, and the second pass would have to re-read every page to work out which half it was in. One dispatch, once the direction is known.

Two things settled while waiting, both recorded in the body above:

1. The field is seoTitle, with .trim().min(1) — an empty value fails the build rather than producing a bare | ObjectOS.
2. The body previously mis-cited Sitemap and hreflang advertise seven locales for pages that exist only in English — 41 of 79 pages #169 as the card covering the missing-translation gap. It was not; it was the sitemap/hreflang card and it has merged. That gap has no card today.

---

Generated by Claude Code