Weave Code
Ai Maria Db Store Laravel Package
symfony/ai-maria-db-store
MariaDB vector store integration for Symfony AI Store, enabling vector search and RAG with MariaDB's native VECTOR type, VECTOR INDEX, and distance functions. Requires MariaDB 11.7 or later.
Provides a MariaDB vector store integration for Symfony AI Store, letting you persist and query embeddings directly in your existing MariaDB database. It enables RAG (Retrieval-Augmented Generation) workflows by storing vectors alongside your relational data and performing similarity searches using native distance functions.
Requires MariaDB >= 11.7 for vector support.
Features:
- Native
VECTORcolumn storage for embeddings VECTOR INDEXsupport for fast similarity search- Distance functions for nearest-neighbor queries
- Seamless integration with Symfony AI Store
- Documented RAG setup and vector overview links
Frequently asked questions about Ai Maria Db Store
- What is symfony/ai-maria-db-store used for?
- It provides a MariaDB vector store integration for Symfony AI Store. It enables vector similarity search and RAG workloads on MariaDB 11.7+ using native VECTOR columns, vector indexes, and distance functions. It is a good fit for database-native, cost-sensitive architectures rather than GPU-accelerated or very high-QPS vector workloads.
- Does symfony/ai-maria-db-store work with Laravel out of the box?
- No. The package is Symfony-centric and is not a native Laravel package. Laravel integration requires bridges, facades, custom service providers, configuration mapping, or raw PDO usage. You should isolate the store behind your own interface to reduce coupling.
- How do I install symfony/ai-maria-db-store for a Laravel project?
- Install it with Composer and confirm your PHP and Symfony AI dependency versions satisfy package v0.14.0 requirements. You also need MariaDB 11.7+ for native vector support. Because Laravel does not provide the Symfony service container wiring automatically, plan to add a bridge or custom service provider.
- What MariaDB version is required for vector search?
- MariaDB 11.7 or newer is mandatory for native vector support. Earlier MariaDB versions do not provide the VECTOR column, vector index, and distance function features this package relies on. Upgrading existing deployments may introduce operational disruption, so validate in staging first.
- How do I configure the MariaDB vector store in Laravel?
- There is no native Laravel YAML-to-config mapping, so you must map Symfony AI store configuration into Laravel config, such as config/ai.php, and bootstrap the store through a service provider or bridge. VECTOR columns, indexes, and schema evolution remain manual operational work. Keep raw SQL vector queries isolated from Eloquent to avoid query builder interference.
- Can I use the count capability for pagination and RAG dashboards?
- Version 0.14.0 adds a store count capability that is useful for pagination, admin dashboards, RAG context budgeting, adoption metrics, and monitoring. Confirm whether your implementation returns exact or approximate counts and whether the query is index-assisted. Benchmark count performance on large vector tables because full scans can be expensive.
- What is VectorDocumentInterface and why does it matter?
- VectorDocumentInterface standardizes vector document handling and improves abstraction for Symfony AI stores. It reduces future swap risk if you isolate the store behind an interface. Existing Laravel wrappers, DTOs, or repository mappings that assumed concrete document types may need adaptation.
- Is symfony/ai-maria-db-store suitable for high-QPS production vector search?
- It is CPU-bound and does not provide GPU acceleration, so latency can become a concern for large datasets or high QPS. Horizontal scaling still requires manual sharding or read replicas, and distance metrics are limited to common options such as cosine, Euclidean, and L2. It fits cost-sensitive, database-native architectures better than very high-throughput vector workloads.
- How should I test MariaDB vector store integration in Laravel?
- Use a MariaDB 11.7+ test database with real VECTOR columns and vector indexes rather than mocking all SQL behavior. Write integration tests for ingestion, distance queries, filtering, and count behavior, and keep store access behind a repository interface. Also verify the exact Symfony AI version required by package v0.14.0 in your test matrix.
- What alternatives or maintenance risks should I consider?
- The package is still 0.x, so minor releases may include contract-affecting changes even when no explicit breaking changes are listed. VectorDocumentInterface and the new count capability can cause short-term integration churn if your wrappers expose concrete document types or store capabilities directly. If you need GPU acceleration, high QPS, or native Laravel support, evaluate dedicated vector databases or Laravel-native alternatives before committing.
Weaver
How can I help you explore Laravel packages today?