Anvia

Appearance

@anvia/redis ​

@anvia/redis stores embedded documents as Redis hashes and searches them through a RediSearch HNSW vector index.

Install ​

sh
pnpm add @anvia/redis @anvia/core @anvia/openai redis

The ESM package includes the Redis client and peers with @anvia/core >=0.7.1 <1.0.0. The target Redis deployment must support the FT.* search commands used by the adapter.

Store and search documents ​

ts
import { embedDocuments } from '@anvia/core/embeddings'
import { RedisVectorStore } from '@anvia/redis'
import { OpenAIClient } from '@anvia/openai'

const openai = new OpenAIClient({
  apiKey: process.env.OPENAI_API_KEY,
})
const embeddings = openai.embeddingModel('text-embedding-3-small')
const sourceDocuments = [
  {
    id: 'password-reset',
    text: 'Password reset links expire after 30 minutes.',
  },
]

const documents = await embedDocuments(embeddings, sourceDocuments, {
  id: (document) => document.id,
  content: (document) => document.text,
})

const store = await RedisVectorStore.connect({
  indexName: 'support_docs',
  keyPrefix: 'knowledge:support:',
  vectorSize: 1536,
})

await store.upsertDocuments(documents)

const results = await store.index(embeddings).search({
  query: 'How do I reset a password?',
  topK: 5,
})

Without an injected client, the adapter connects to REDIS_URL or redis://localhost:6379.

Index ownership ​

By default, connect() checks the index and creates a hash-backed HNSW index when missing. The default key prefix is anvia:<indexName>: and the default distance is COSINE. The vector field uses FLOAT32 with the configured dimension.

For production, create and tune the search index through deployment automation, then pass createIfMissing: false. Keep indexName, prefix, dimension, distance, and field layout consistent with the adapter.

The adapter writes hashes without expiration. Retention, deletion, and stale-document cleanup belong to the application. Metadata names beginning with __anvia_ are reserved.

Production patterns ​

  • Inject an already-connected client when the application owns reconnect and shutdown behavior.
  • Use a dedicated, collision-free key prefix.
  • Plan memory capacity for serialized documents and vectors.
  • Provision RediSearch and validate index readiness before application startup.
  • Add an explicit corpus replacement or garbage-collection process.

Reference ​

Built for Anvia.