Statamic Auto Alt Text Laravel Package

Product Decisions This Supports

• Version-compatibility / upgrade-risk decisions. The headline change in v2.0.5 is a Composer constraint fix: the package now accepts Prism 0.100 alongside 0.99 instead of pinning to a single 0.x minor. In practice this decides whether you can install this package at the same time as anything else in your dependency tree that requires Prism 0.100. If you evaluated this package and walked away because of a conflict, re-check it — that specific blocker is gone.
• Build vs. buy for LLM plumbing. This is a thin integration layer over Prism (the PHP LLM client), not a standalone AI platform. That makes it a low-commitment "buy" for teams that want Prism's provider abstraction without hand-rolling the Laravel wiring. Your exit cost is roughly "delete the package and call Prism directly," which is a much easier committee conversation than adopting a proprietary AI SDK.
• Roadmap sequencing around LLM features. Because Prism minors now arrive as range-widening PRs rather than hard pins, upstream provider/model support (new models, new providers) should reach you without a coordinated multi-package upgrade. That de-risks putting AI features on a near-term roadmap: you're not gated on this package shipping a release every time Prism does.
• Maintenance-health as a selection criterion. The CI now runs against every supported Prism minor with a dedicated test over the Prism request chain, and the Dependabot config was repaired (an unsupported cooldown option had been silently disabling all version updates). Two signals here: the maintainer responds to real-world install friction (issues #17/#18), and the automation that keeps the package current was actually broken and is now fixed. Both are positive, but they're also a reminder to verify dependency automation isn't quietly dead in any package you depend on.
• Use cases this fits. Laravel applications adding chat/completion/embedding features, internal tooling and prototypes where speed matters more than vendor-specific optimization, and teams already standardized on Prism who want the Laravel-side ergonomics.

When to Consider This Package

Consider it when:

• You're on Laravel and already using, or willing to adopt, Prism as your LLM client — this package assumes Prism, it doesn't replace it.
• You need to coexist with other packages that pull Prism 0.100; v2.0.5 is the release that makes that possible, so pin ^2.0.5 rather than ^2.0.
• Your team values a small, auditable integration layer over a batteries-included AI framework.
• You want upstream model/provider churn to arrive via automated range-widening PRs instead of manual upgrade projects.

Look elsewhere when:

• You need a stable, post-1.0 Prism API to build against. Prism is still on 0.x, which is why caret constraints pin the minor and why this class of breakage exists at all. Expect more of these constraint widenings as Prism approaches 1.0 — budget for it.
• Your requirements are provider-specific (fine-grained control over a single vendor's features, native SDK capabilities, prompt-caching or batch APIs that a thin abstraction may not surface). Verify actual capability coverage against the README before committing; the release notes here tell you about packaging, not features.
• You're not on Prism and don't intend to be — the dependency is the architecture.
• You need enterprise support, an SLA, or a vendor to call. This is community-maintained; the responsiveness shown in #18/#19 is good faith, not a contract.

Note: v2.0.5 is a patch release with no new user-facing features and no deprecations. It changes installability and CI hygiene only. No migration work is required beyond a composer update.

How to Pitch It (Stakeholders)

For executives (one paragraph): We're evaluating a small open-source Laravel component that connects our application to large-language-model providers via Prism, an established PHP AI client. It's a thin layer, not a platform bet — adopting it doesn't lock us to a vendor, and removing it later means calling Prism directly rather than a rewrite. The new v2.0.5 release is maintenance-only: it removes a version conflict that prevented installation alongside other packages using the current Prism release, and it adds automated testing across every Prism version we'd support. No breaking changes, no migration cost.

For engineering (bullets):

• Prism constraint widened from ^0.99 to accept 0.100 too — the previous pin was a Composer 0.x caret artifact, not intentional.
• CI now matrix-tests against every supported Prism minor, with a new test covering the Prism request chain — so Prism breakage surfaces in CI rather than in production.
• Dependabot was misconfigured (unsupported cooldown key in the github-actions block disabled all version updates); now fixed, Composer updates use the widen strategy so future Prism minors land as non-breaking range-widening PRs.
• Upgrade is a patch bump; require ^2.0.5. Confirm your lockfile isn't still resolving 2.0.4 or older.
• Residual risk: upstream Prism is pre-1.0. Expect further minor-pinning friction until it stabilizes; that risk lives with Prism, not with this package's code.