Anvia

@anvia/chroma ​

@anvia/chroma stores Anvia embedded documents in ChromaDB and exposes them through the SDK's vector-search interface.

Install ​

sh
pnpm add @anvia/chroma @anvia/core @anvia/openai chromadb

The package is ESM-only, includes chromadb, and should use a version compatible with its declared @anvia/core dependency range.

Store and search documents ​

ts
import { retrieveDocuments } from "@anvia/core/vector-store";
import { embedDocuments } from '@anvia/core/embeddings';
import { ChromaVectorClient } from '@anvia/chroma';
import { OpenAIClient } from '@anvia/openai';
const openai = new OpenAIClient({
    apiKey: process.env.OPENAI_API_KEY!,
});
const embeddings = openai.embeddingModel({
    modelId: 'text-embedding-3-small'
});
const sourceDocuments = [
    {
        id: 'password-reset',
        text: 'Password reset links expire after 30 minutes.',
        category: 'account',
    },
];
const { documents } = await embedDocuments({
    model: embeddings,
    documents: sourceDocuments,
    id: (document) => document.id,
    content: (document) => document.text,
    metadata: (document) => ({ category: document.category })
});
const storeClient = new ChromaVectorClient({});
const store = storeClient.vectorStore({
    collectionName: 'support_docs',
    dimensions: 1536
});
await store.ensure();
await store.upsert({
    documents: documents
});
const results = await retrieveDocuments({
    store: store,
    model: embeddings,
    query: 'How do I reset a password?',
    topK: 5
});

The store implements the SDK's VectorStore interface: ensure(), validate(), upsert(), and search(). To expose retrieval as an agent tool, build one with createVectorSearchTool() from @anvia/core/vector-store. See Vector stores and Search tools.

Collection ownership ​

ensure() gets or creates the collection. It supplies embeddingFunction: null because Anvia writes precomputed embeddings, and unless the store was created with a custom configuration, it requests an HNSW index in the store's distance space (cosine by default). validate() only checks an existing collection.

For production, provision the collection with your infrastructure workflow and use:

ts
const storeClient = new ChromaVectorClient({
    client
});
const store = storeClient.vectorStore({
    collectionName: 'support_docs',
    dimensions: 1536
});
await store.validate();

Pass metadata or configuration when Studio should create the collection with explicit Chroma settings. Keep those settings compatible with the embedding model used to produce stored vectors.

Production patterns ​

  • Inject an authenticated, application-managed Chroma client instead of relying on the no-argument default client.
  • Keep collection creation in deployment infrastructure and fail startup when it is missing.
  • Preserve stable document IDs so repeated ingestion updates the same vector records.
  • Reserve capacity and retention around the number of embeddings, not only source documents.
  • Apply metadata filters for tenant or corpus boundaries; filters are not a substitute for authorization.

Reference ​

Built for Anvia.