Ai Bundle Laravel Package

Product Decisions This Supports

• Framework glue build vs. buy: The bundle is the "buy" option for Symfony dependency-injection wiring, config schema validation, and service autoconfiguration around the Lingoda AI PHP SDK. It saves roughly one to two days of boilerplate per Symfony service — a real but small win, and one we could replicate ourselves.
• AI abstraction layer decision: This bundle only matters if we have already decided the Lingoda AI PHP SDK is our provider abstraction. If that decision is still open, this package should not drive it; it is downstream of it.
• Roadmap scoping by stack: Relevant only for Symfony-based services. For a Laravel-first portfolio, this is a "no" on the current roadmap rather than a "maybe" — see the next section.
• Feature work it implies but does not guarantee: Environment-based provider/model switching, externalized prompt and credential config, and centralized AI call configuration. Confirm these exist in the underlying SDK before crediting the bundle with them.
• Use case fit: Internal tools, back-office AI features, and prototypes where we want config-driven model selection without hand-rolling a wrapper.
• Dependency risk posture: One star, zero dependents, single vendor, no visible BC policy — this is an internal extraction published publicly, not a community-vetted library. The portfolio decision shifts from "adopt" to "vendor/fork or use as reference code."

When to Consider This Package

Consider it when:

• We are already on Symfony and the Lingoda AI PHP SDK is our chosen AI abstraction.
• We want framework-native YAML config, DI autowiring, and env-based provider/model switching without writing our own service provider.
• We accept owning the upgrade path and are willing to pin the version and vendor the source.
• The team is small enough that a day of glue-code savings is genuinely material.

Look elsewhere / do not adopt when:

• We are on Laravel. This is a Symfony bundle. Laravel gets no package discovery, no bundle config, no container wiring from it — we would write a service provider against the SDK regardless. The brief framed this as a Laravel/PHP package; it is Symfony-only, and that is the decisive fact.
• We need community support, a security review process, CVE handling, or multiple maintainers. One star and zero dependents fails most dependency-review thresholds.
• We need documented backward-compatibility guarantees; a single 2026-10-14 release with no visible cadence or changelog history is a planning risk.
• We run multiple frameworks (Symfony plus Laravel) and need one framework-agnostic layer.
• We need production-grade observability, token accounting, retries, circuit-breaking, or eval hooks — verify these live in the SDK rather than assuming the bundle supplies them.

How to Pitch It (Stakeholders)

Executives: This is a small, free (MIT) adapter that plugs an AI SDK into Symfony's configuration and dependency-injection system. It saves roughly a day of engineering per Symfony service and standardizes how we call AI models across applications. It is not a platform: one maintainer, one star, zero known dependents. Recommendation: use it only where we are already standardized on the Lingoda AI SDK and Symfony; otherwise treat it as reference code and build the thin wrapper ourselves.

Engineering: The bundle provides Symfony DI wiring, a config schema, and autoconfiguration for the Lingoda AI PHP SDK — nothing more, and nothing for Laravel. Because it is single-vendor with low adoption and no visible BC policy, if we adopt it we pin the version, vendor the source, and keep our own thin interface in front of the SDK so that swapping providers or replacing the integration layer stays a contained change.