Sulu Translations Bundle Laravel Package

Product Decisions This Supports

• Localization roadmap: 0.4.0 activates Sulu's AI translation feature for translation keys, turning multi-locale content rollout from a manual translator-dependent workflow into a largely automated one. This directly informs decisions about how many markets/locales we can realistically support next quarter and whether we staff translation vendors or lean on AI-assisted drafts with human review.
• Build vs. buy on translation tooling: Rather than building a bespoke translation management integration, this bundle is a thin, opinionated layer on top of Sulu's native AI translation capability. If Sulu is already the CMS, this is close to a "buy" decision with minimal integration cost.
• Content velocity metrics: AI-assisted key translation reduces per-locale turnaround time — a metric worth tracking as a leading indicator for market-entry experiments and A/B tests across locales.
• Scope decision for the i18n epic: Because the feature lands on translation keys (static UI/string resources) rather than full content pages, it cleanly separates "translate the interface" from "translate the editorial content." That boundary helps sequence the roadmap: interface localization first, content translation later.
• Risk/quality gate: AI-generated translations require a review workflow. This release supports a decision to keep human sign-off in the loop for customer-facing or legally sensitive strings, and to auto-accept for internal/admin surfaces.

When to Consider This Package

Adopt when:

• The product is built on Sulu CMS (Symfony-based). Note this is not a Laravel package — the original framing was off; if the stack is Laravel, this bundle is not applicable and you should look at Laravel-native options (laravel-lang, astrotomic/laravel-translatable, or a Spatie/DeepL translation pipeline).
• You manage translations as key-based string resources and want AI drafts generated in-place inside the CMS admin rather than in a separate TMS.
• You need a low-ceremony, single-vendor implementation and are comfortable with a small, early-stage package.
• You want the option to fall back to manual translation for specific keys — the bundle augments rather than replaces the existing workflow.

Look elsewhere when:

• The stack is Laravel or plain PHP without Sulu — the package is not portable.
• You require enterprise TMS features: translation memory, glossary enforcement, vendor management, per-string audit trails, or SLA-backed support. A dedicated TMS (Phrase, Lokalise, Crowdin) integrated via API is the safer call.
• You need content/document translation rather than key translation — out of scope here.
• Compliance or regulated content requires a certified human translation pipeline with no AI in the loop.
• You are unwilling to depend on a 0.x, single-maintainer-adjacent package for a production-critical path.

Maturity caveat: at 0.4.0 with a first-time contributor landing the headline feature, treat this as early-stage. Pin versions, wrap the integration behind an internal interface, and budget for the possibility of maintaining a fork.

How to Pitch It (Stakeholders)

To executives: This release adds AI-assisted translation of interface strings directly inside our CMS. In practical terms it lowers the cost and calendar time to launch each additional language — we can go from "months of translator coordination" to "days of AI drafting plus a review pass." It's a small dependency, not a platform bet, and it de-risks our multi-market expansion plans without new vendor contracts. Key caveat: it covers interface text, not marketing or editorial content, and it's an early-stage 0.x package, so we'll pin the version and keep a review step for anything customer-facing.

To engineering: Version 0.4.0 activates Sulu's built-in AI translation for translation keys. Integration surface is small — it hooks into Sulu's existing translation admin rather than introducing a new service. Two asks: (1) confirm the AI provider configuration, credential handling, and data-residency implications of sending strings to a third-party model; (2) wrap the bundle behind a thin internal abstraction so a swap to a TMS API is a contained change. Also note this is a Sulu/Symfony bundle — it has no relevance if any part of the localization effort moves to our Laravel services, which will need a separate solution.

Recommendation: Approve for a piloted rollout on one non-critical locale, gated on a review workflow for customer-facing strings. Revisit for wider adoption once the package reaches 1.0 or after we've run it for a full release cycle without incident.