Weave Code
Ai Chroma Db Store Laravel Package
symfony/ai-chroma-db-store
ChromaDB vector store integration for Symfony AI Store, enabling collection management and vector queries against ChromaDB.
ChromaDB vector store bridge for Symfony AI
Frequently asked questions about Ai Chroma Db Store
- How do I install symfony/ai-chroma-db-store in a Laravel app?
- Install it with Composer using the package name symfony/ai-chroma-db-store, and verify the exact symfony/ai constraint in the package composer.json before choosing versions. There is no Laravel service provider or config publishing, so you will still bind StoreInterface manually through StoreFactory in a service provider. Make sure your ChromaDB endpoint is reachable before wiring the store into application services.
- Does symfony/ai-chroma-db-store support Laravel out of the box?
- No. It is a Symfony AI Store bridge for ChromaDB, not a Laravel package, and it ships no bundle, autoconfiguration, or Laravel-specific service provider. Laravel integration is manual: bind StoreInterface via StoreFactory and provide environment-driven configuration for host, port, TLS, token, tenant, database, collection, metric, and timeout.
- Which symfony/ai version pairs with symfony/ai-chroma-db-store v0.14?
- The store package tracks the Symfony AI monorepo release train, so v0.14 of the store pairs with v0.14 of symfony/ai. Upgrade both in lockstep rather than half-upgrading one side, because interface mismatches can fail at container boot instead of at first use. Always check the exact symfony/ai constraint in the store package composer.json before estimating an upgrade.
- How do I implement VectorDocumentInterface for a Laravel model or DTO?
- Implement VectorDocumentInterface directly on your domain DTO or model instead of building against the concrete Document class shown in older examples. That collapses the document-mapping layer and makes test doubles much simpler. It also reduces exposure to upstream constructor changes, but only if you type-hint and build against the interface from day one.
- Can I use count() to reconcile ChromaDB vectors with database rows?
- Yes. Version 0.14 adds count() on stores, which lets you compare source-of-truth rows with vectors in a collection as a one-line health metric or nightly reconciliation job. Previously this required a query hack or a full scan. Use it in the pilot rather than deferring it, because it closes the cheapest observability gap.
- How do I configure ChromaDB host, port, token, and collection in Laravel?
- There is no config publishing, so create your own Laravel config or environment variables and pass values into StoreFactory when binding StoreInterface. Cover host, port, TLS, token, tenant, database, collection, metric, and timeout, and include an authenticating health check. Keep this binding small but permanent, because it is glue you own.
- Does symfony/ai-chroma-db-store create ChromaDB collections or set HNSW parameters?
- No. Collection lifecycle remains unmanaged by the package: creating collections, choosing the distance metric, and setting HNSW parameters are your responsibility. Version collection names and treat them as schema. Provision collections outside the store before relying on queries or count() in production.
- Who handles embeddings and vector dimensions with this ChromaDB store?
- The package is a vector store bridge, not an embedding pipeline, and it has no embedding or Vectorizer story. You own provider choice, model version, vector dimensions, and re-embedding cost, then pass vectors into ChromaDB through the store. The embedding pipeline is typically the largest hidden cost in this integration.
- How should I test a Laravel integration with this ChromaDB vector store?
- Use VectorDocumentInterface to create simple test doubles, and bind a fake or mock StoreInterface through StoreFactory in your test service provider. Add at least one integration test against an ephemeral ChromaDB instance for health checks and count(). Re-baseline retrieval fixtures before comparing quality, because queries without an explicit limit now inherit ChromaDB's server-side default result-set size.
- Is symfony/ai-chroma-db-store stable enough for production in Laravel?
- Treat it as a time-boxed spike rather than an immediate roadmap commitment. The package is pre-1.0 and has moved from v0.8.x to v0.14.x with behavior changes, including a new interface and query-limit behavior, so pin exact versions and treat every bump as a change request. It remains a narrow, auditable bridge, but Laravel still lacks queue integration, observability, and framework wiring.
Weaver
How can I help you explore Laravel packages today?