Weaviate

Weaviate is a cloud-native, open-source vector database with built-in support for hybrid search and multi-tenancy. synaptic-rag (with feature weaviate) implements the [VectorStore] trait using the Weaviate v1 REST API.

Setup

[dependencies]
synaptic = { version = "0.4", features = ["weaviate"] }

Run a local Weaviate instance with Docker:

docker run -d -p 8080:8080 -p 50051:50051 cr.weaviate.io/semitechnologies/weaviate:latest

Or use Weaviate Cloud Services.

Configuration

use synaptic::weaviate::{WeaviateVectorStore, WeaviateConfig};

// Local Weaviate
let config = WeaviateConfig::new("http", "localhost:8080", "Documents");

// Weaviate Cloud Services (WCS) with API key
let config = WeaviateConfig::new("https", "my-cluster.weaviate.network", "Documents")
    .with_api_key("wcs-secret-key");

let store = WeaviateVectorStore::new(config);

// Create class schema (idempotent — safe to call multiple times)
store.initialize().await?;

Configuration reference

Field	Type	Default	Description
`scheme`	`String`	required	`"http"` or `"https"`
`host`	`String`	required	Host and port (e.g. `localhost:8080`)
`class_name`	`String`	required	Weaviate class name (must start with uppercase)
`api_key`	`Option<String>`	`None`	API key for WCS authentication

Adding documents

use synaptic::weaviate::{WeaviateVectorStore, WeaviateConfig};
use synaptic::core::Document;
use synaptic::openai::OpenAiEmbeddings;
use std::sync::Arc;

let config = WeaviateConfig::new("http", "localhost:8080", "Articles");
let store = WeaviateVectorStore::new(config);
store.initialize().await?;

let embeddings = Arc::new(OpenAiEmbeddings::new(/* config */));

let docs = vec![
    Document::new("1", "Rust is a systems programming language."),
    Document::new("2", "Weaviate is a vector database."),
    Document::new("3", "Synaptic is a Rust agent framework."),
];

let ids = store.add_documents(docs, embeddings.as_ref()).await?;
println!("Added {} documents", ids.len());

Similarity search

use synaptic::core::VectorStore;

let results = store.similarity_search("systems programming", 3, embeddings.as_ref()).await?;
for doc in results {
    println!("[{}] {}", doc.id, doc.content);
}

Deleting documents

store.delete(&["weaviate-uuid-1".to_string(), "weaviate-uuid-2".to_string()]).await?;

RAG pipeline

use synaptic::retrieval::VectorStoreRetriever;
use synaptic::core::Retriever;
use std::sync::Arc;

let store = Arc::new(WeaviateVectorStore::new(config));
let retriever = VectorStoreRetriever::new(store, embeddings, 4);

let docs = retriever.get_relevant_documents("Rust async programming").await?;

Error handling

use synaptic::core::SynapticError;

match store.similarity_search("query", 5, embeddings.as_ref()).await {
    Ok(docs) => println!("Found {} results", docs.len()),
    Err(SynapticError::VectorStore(msg)) => eprintln!("Weaviate error: {msg}"),
    Err(e) => return Err(e.into()),
}

Class schema

initialize() creates the following Weaviate class if it does not exist:

{
  "class": "Documents",
  "vectorizer": "none",
  "properties": [
    { "name": "content",  "dataType": ["text"] },
    { "name": "docId",    "dataType": ["text"] },
    { "name": "metadata", "dataType": ["text"] }
  ]
}

Vectors are provided by Synaptic (no Weaviate vectorizer module needed).

Keyboard shortcuts

Synaptic