Lithuanian Names Declension Laravel Package

Product Decisions This Supports

• Lithuanian-market localization roadmap. Declension (linksniavimas) is the gate between "translated" and "native-sounding" Lithuanian. A PM scoping an LT locale for onboarding, checkout, or CRM will hit the problem that names change form by grammatical case — Jonas becomes Jonai in a greeting, Jono in a possessive, Jonui for "to Jonas." This package is a small piece of that puzzle (typically nominative → genitive/dative/vocative for first and last names).
• Personalization quality in transactional surfaces. The highest-value use cases are cheap to enumerate: greeting lines in email/SMS/push (Sveiki, Jonai vs. the wrong Sveiki, Jonas), invoice and contract headers, shipping labels, "assigned to" fields in a CRM, and formal B2B salutations. Wrong declension is not a bug users report; it is a trust signal they silently downgrade.
• Build vs. buy vs. copy. This is a ~10-year-old, 3-star, zero-dependent bundle. The right decision is not "adopt a dependency" but "treat it as a specification and a test corpus." Use it to enumerate the rules and edge cases (consonant vs. vowel stems, -as/-is/-ys/-us/-a/-ė/-ė endings, foreign names that should not be declined, hyphenated surnames), then implement or fork 200–400 lines behind your own interface. Lithuanian morphology is rule-based enough that a maintained internal class beats an unmaintained vendor package.
• Copy strategy as an alternative to code. Before building anything, decide whether the product can sidestep declension: many strings can be rewritten to avoid a declined name (Sveiki! alone, or the name in a separate field/label). That is a zero-dependency fix and should be evaluated first — it often removes 60–80% of the surface area.
• Reverse direction / data hygiene. The same rule set supports normalizing a declined form back to nominative for deduplication and CRM record matching — worth a spike if name identity matters in your data model.

When to Consider This Package

Consider it (as reference material, not a runtime dependency) when:

• Lithuanian is a shipped or committed locale, and names appear inside sentence-level copy rather than in isolated labeled fields.
• You have a bounded set of name fields (first name, last name) and formal-correspondence templates you can test against.
• You accept owning the code: vendor it or fork it, pin it, add tests, and wrap it behind your own NameDecliner interface so the implementation can be swapped.

Look elsewhere — or skip — when:

• You do not serve the Lithuanian market, or you serve it only in nominative contexts (name in a standalone field, avatar, list row).
• You need a maintained dependency: last release 2015, no recent Laravel auto-discovery, likely PHP 5-era constraints, no security process, no issue response, and 3 stars means no community safety net. Do not put this in require on a modern Laravel 10/11 app without a compatibility spike.
• You need genuine morphology beyond first/last names — place names, addresses, organization names, or sentence-level inflection. That is an NLP problem, and a 2015 rule table is not it.
• Correctness is contractual (legal documents, government filings). Buy a vetted Lithuanian language tool or use a human review step instead of an unmaintained library.
• Compliance or provenance review flags unpinned third-party code with no maintainer.

How to Pitch It (Stakeholders)

To executives: This is not a platform decision — it is a small, MIT-licensed Lithuanian grammar utility with essentially no adoption and no maintenance since 2015. The opportunity is not the package; it is the localization gap it exposes. If Lithuania is on the roadmap, we should budget roughly one engineering week to own a name-declension component (forked or written in-house from this package's rules) plus a copy review, rather than take a supply-chain dependency. The payoff is small but real: greetings, invoices, and notifications that read as native Lithuanian instead of obviously machine-translated. If Lithuania is not on the roadmap, close this out — the correct action is "no."

To engineering: Read it, don't require it. Extract the declension rules and the exception list into our own tested class behind an interface, validate against a fixture set of Lithuanian, foreign, and hyphenated names, and keep a fallback that renders the nominative form safely when confidence is low. Estimate: 1–2 days for a spike, 1 week including tests and template integration. Flag the PHP-version and Laravel-compatibility gap as the reason we fork rather than install.