---
title: "Cloud Vector Backends"
description: "Cloud vector backends provide scalable, managed infrastructure for high-dimensional vector persistence and approximate nearest neighbor (ANN) retrieval across distributed cloud storage engines and ..."
last_updated: "2026-10-11T17:37:10.218815+00:00"
canonical_url: "https://www.doc0.dev/docs/faa36707-7c28-4f69-a18f-700ff61c704e/technical/vector-stores/cloud-vector-backends"
---

<details>
<summary>Relevant source files</summary>

The following files were used as context for generating this wiki page:

- [mem0/vector_stores/s3_vectors.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/s3_vectors.py)
- [mem0-ts/src/oss/src/vector_stores/s3_vectors.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/s3_vectors.ts)
- [mem0/vector_stores/neptune_analytics.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/neptune_analytics.py)
- [mem0-ts/src/oss/src/vector_stores/neptune_analytics.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/neptune_analytics.ts)
- [mem0/vector_stores/weaviate.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/weaviate.py)
- [mem0/vector_stores/turbopuffer.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/turbopuffer.py)
- [mem0/vector_stores/upstash_vector.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/upstash_vector.py)
- [mem0-ts/src/oss/src/vector_stores/turbopuffer.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/turbopuffer.ts)
- [mem0/vector_stores/valkey.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/valkey.py)
- [mem0/vector_stores/opensearch.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/opensearch.py)
- [mem0-ts/src/oss/src/vector_stores/qdrant.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/qdrant.ts)
- [mem0-ts/src/oss/src/vector_stores/milvus.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/milvus.ts)
- [mem0/vector_stores/__init__.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/__init__.py)
- [mem0-ts/src/oss/src/vector_stores/baidu.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/baidu.ts)
- [mem0/configs/vector_stores/__init__.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/configs/vector_stores/__init__.py)
- [mem0-ts/src/oss/src/vector_stores/databricks.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/databricks.ts)
- [mem0/vector_stores/chroma.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/chroma.py)
- [mem0-ts/src/oss/src/vector_stores/azure_ai_search.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/azure_ai_search.ts)
- [mem0-ts/src/oss/src/vector_stores/oracledb.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/oracledb.ts)
- [mem0-ts/src/oss/src/vector_stores/weaviate.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/weaviate.ts)
- [mem0-ts/src/oss/src/vector_stores/upstash_vector.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/upstash_vector.ts)
- [mem0/vector_stores/langchain.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/langchain.py)
- [mem0/vector_stores/mongodb.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/mongodb.py)
- [mem0-ts/src/oss/src/vector_stores/pinecone.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/pinecone.ts)
- [mem0/vector_stores/milvus.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/milvus.py)
- [mem0/vector_stores/azure_ai_search.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/azure_ai_search.py)
- [mem0/vector_stores/supabase.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/supabase.py)
- [mem0-ts/src/oss/src/vector_stores/vectorize.ts](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/vectorize.ts)
- [mem0/vector_stores/faiss.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/faiss.py)
- [mem0/vector_stores/pinecone.py](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/pinecone.py)
</details>

## Overview

Cloud vector backends provide scalable, managed infrastructure for high-dimensional vector persistence and approximate nearest neighbor (ANN) retrieval across distributed cloud storage engines and managed database services. These storage backends form the core persistence layer for memory management systems, handling vector insertions, multi-tenant namespace isolation, metadata filtering, and hybrid sparse-dense keyword searches. By abstracting diverse remote protocols—ranging from object-backed S3 endpoints and serverless HTTP APIs to graph analytics engines and distributed index clusters—these drivers enable uniform vector operations across both Python and TypeScript runtimes.

Sources: [mem0/vector_stores/s3_vectors.py:24-181](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/s3_vectors.py#L24-L181), [mem0-ts/src/oss/src/vector_stores/s3_vectors.ts:35-242](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/s3_vectors.ts#L35-L242), [mem0/vector_stores/neptune_analytics.py:41-199](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/neptune_analytics.py#L41-L199), [mem0-ts/src/oss/src/vector_stores/neptune_analytics.ts:42-240](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/neptune_analytics.ts#L42-L240), [mem0/vector_stores/turbopuffer.py:25-207](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/turbopuffer.py#L25-L207), [mem0-ts/src/oss/src/vector_stores/turbopuffer.ts:13-212](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/turbopuffer.ts#L13-L212), [mem0-ts/src/oss/src/vector_stores/qdrant.ts:300-412](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/qdrant.ts#L300-L412), [mem0/vector_stores/milvus.py:33-146](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/milvus.py#L33-L146), [mem0-ts/src/oss/src/vector_stores/milvus.ts:272-335](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/milvus.ts#L272-L335), [mem0/vector_stores/pinecone.py:25-121](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/pinecone.py#L25-L121), [mem0-ts/src/oss/src/vector_stores/pinecone.ts:244-289](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/pinecone.ts#L244-L289), [mem0/vector_stores/weaviate.py:31-157](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/weaviate.py#L31-L157), [mem0-ts/src/oss/src/vector_stores/weaviate.ts:31-237](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/weaviate.ts#L31-L237), [mem0/vector_stores/upstash_vector.py:41-172](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/upstash_vector.py#L41-L172), [mem0-ts/src/oss/src/vector_stores/upstash_vector.ts:17-268](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/upstash_vector.ts#L17-L268)

## Cloud Vector Driver Abstractions

### Architectural Contract and Client Initialization

Cloud vector storage drivers implement a standard interface across both Python and TypeScript runtimes, wrapping diverse remote SDKs like `boto3`, `@aws-sdk/client-s3vectors`, and `weaviate-client`. Initialization routines handle optional peer dependencies dynamically, establish remote connections across HTTP, gRPC, or custom client parameters, and provision underlying vector buckets, indices, or classes.

Sources: [mem0/vector_stores/s3_vectors.py:24-41](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/s3_vectors.py#L24-L41), [mem0-ts/src/oss/src/vector_stores/s3_vectors.ts:35-123](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/s3_vectors.ts#L35-L123), [mem0/vector_stores/weaviate.py:31-85](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/weaviate.py#L31-L85), [mem0-ts/src/oss/src/vector_stores/weaviate.ts:31-118](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/weaviate.ts#L31-L118)

> [!NOTE]
> Optional SDK peer dependencies such as `@aws-sdk/client-s3vectors` and `weaviate-client` are lazily imported at runtime in TypeScript and safely caught behind missing-import exceptions in Python to keep base package installations lightweight for consumers who exclude specific cloud backends.

Sources: [mem0/vector_stores/s3_vectors.py:9-13](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/s3_vectors.py#L9-L13), [mem0-ts/src/oss/src/vector_stores/s3_vectors.ts:69-82](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/s3_vectors.ts#L69-L82), [mem0/vector_stores/weaviate.py:9-13](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/weaviate.py#L9-L13), [mem0-ts/src/oss/src/vector_stores/weaviate.ts:49-59](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/weaviate.ts#L49-L59)

### Lifecycle Hooks and Data Operations

Drivers manage the complete entity lifecycle across insertion, vector search, metadata filtering, retrieval, updates, and deletion. Depending on the engine capabilities, updates are handled either via direct object overwrites, native upsert batches, or pre-flight fetching to resolve missing vector payloads.

| Operation | Python Method | TypeScript Method | Core Responsibility |
| :--- | :--- | :--- | :--- |
| **Insert** | `insert(...)` | `insert(...)` | Batch writing of vectors, IDs, and metadata payloads into indices or collections. |
| **Search** | `search(...)` | `search(...)` | Approximate nearest neighbor vector query with optional metadata filters and top-$k$ limits. |
| **Get** | `get(...)` | `get(...)` | Retrieval of stored vector record details and metadata by primary identifier. |
| **Update** | `update(...)` | `update(...)` | Modification of vector data or associated payload properties, handling missing parameters via pre-flight lookups. |
| **Delete** | `delete(...)` | `delete(...)` | Removal of individual vector entries by identifier. |

Sources: [mem0/vector_stores/s3_vectors.py:98-181](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/s3_vectors.py#L98-L181), [mem0-ts/src/oss/src/vector_stores/s3_vectors.ts:125-242](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/s3_vectors.ts#L125-L242), [mem0/vector_stores/weaviate.py:158-199](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/weaviate.py#L158-L199), [mem0-ts/src/oss/src/vector_stores/weaviate.ts:128-204](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/weaviate.ts#L128-L204)

> [!WARNING]
> When executing metadata-only updates in S3 Vectors where `vector=None`, the driver must fetch the existing vector payload beforehand; passing `None` directly to `put_vectors` triggers parameter validation errors in the underlying SDK.

Sources: [mem0/vector_stores/s3_vectors.py:136-168](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/s3_vectors.py#L136-L168)

## S3 Vectors Storage Architecture

The S3 Vectors driver provisions and interacts with object-backed storage collections across Python and TypeScript runtimes using `boto3` and `@aws-sdk/client-s3vectors`. Initialization ensures that both the target vector bucket and the specified index exist before executing any operations, catching `NotFoundException` errors from remote service calls to trigger bucket and index creation automatically.

```mermaid
flowchart TD
    Init[Initialize S3Vectors] --> CheckBucket{Bucket Exists?}
    CheckBucket -- No --> CreateBucket[Create Vector Bucket]
    CheckBucket -- Yes --> CheckIndex{Index Exists?}
    CreateBucket --> CheckIndex
    CheckIndex -- No --> CreateIndex[Create Index with Dimension & Distance]
    CheckIndex -- Yes --> Ready[Store Ready for Operations]
    CreateIndex --> Ready
```

Sources: [mem0/vector_stores/s3_vectors.py:39-70](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/s3_vectors.py#L39-L70), [mem0-ts/src/oss/src/vector_stores/s3_vectors.ts:116-123](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/s3_vectors.ts#L116-L123)

Bucket and index provisioning operations configure data types, vector dimensions, and distance metrics. The supported index configurations and initialization parameters handle dimensional assertions across runtimes.

| Configuration Parameter | Python Type | TypeScript Type | Default Value | Purpose |
| :--- | :--- | :--- | :--- | :--- |
| `vectorBucketName` | `str` | `string` | *Required* | Target S3 bucket container name for vector objects. |
| `collectionName` / `collectionName` | `str` | `string` | *Required* | Index or collection namespace within the vector bucket. |
| `embeddingModelDims` / `dimension` | `int` | `number` | *Required* | Dimensionality of the stored vector embeddings. |
| `distanceMetric` | `str` | `"cosine" \| "euclidean"` | `"cosine"` | Distance calculation metric for similarity searches. |
| `region_name` / `region` | `Optional[str]` | `string` | `undefined` | AWS region hosting the S3 Vectors client. |

Sources: [mem0/vector_stores/s3_vectors.py:25-37](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/s3_vectors.py#L25-L37), [mem0-ts/src/oss/src/vector_stores/s3_vectors.ts:17-29](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/s3_vectors.ts#L17-L29), [mem0-ts/src/oss/src/vector_stores/s3_vectors.ts:47-65](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/s3_vectors.ts#L47-L65)

> [!NOTE]
> Distance score normalization maps raw engine distances into bounded similarity scores. Euclidean distances use an inverse monotonic transformation `1.0 / (1.0 + raw_distance)` since they are unbounded, whereas cosine distances clip `1.0 - raw_distance` at `0.0`.

Sources: [mem0/vector_stores/s3_vectors.py:72-80](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/s3_vectors.py#L72-L80)

## Qdrant and Milvus Ingestion

### Collection Provisioning and Schema Detection

Qdrant and Milvus drivers handle collection provisioning and schema detection differently depending on whether hybrid keyword search is enabled. In Milvus, collection initialization checks whether an existing collection contains both `text` and `sparse` fields, setting an internal `_has_bm25_schema` flag accordingly. When creating a fresh collection, Milvus defines fields for `id` (VARCHAR up to 512 characters, primary key), `vectors` (FLOAT_VECTOR with custom dimensions), `metadata` (JSON), `text` (VARCHAR up to 65535 characters with analyzer enabled), and `sparse` (SPARSE_FLOAT_VECTOR). It then adds a BM25 function linking `text` to `sparse`, applies an `AUTOINDEX` index on `vectors`, and adds a `SPARSE_INVERTED_INDEX` index named `sparse_index` with metric type `BM25`.

Sources: [mem0/vector_stores/milvus.py:53-124](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/milvus.py#L53-L124)

> [!WARNING]
> If an existing Milvus collection predates v3 hybrid search and lacks `text` or `sparse` fields, BM25 keyword scoring is disabled for that collection with a warning logger call, and writing an unknown top-level `text` field will cause insertion rejections.

Sources: [mem0/vector_stores/milvus.py:80-90](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/milvus.py#L80-L90)

### Point Upsert Protocols and Hybrid Handling

Point insertion and upsert workflows adapt to server-side inference capabilities. In Qdrant, point vector building inspects `_has_bm25_slot`. Without a BM25 slot, it returns the plain dense vector array. When a BM25 slot is active, it builds a named vector map containing `""` for the dense vector and a sparse query object for the BM25 vector name using text extracted from `payload.textLemmatized` or `payload.data`.

```typescript
private async upsertPoints(
  points: { id: string | number; vector: any; payload?: any }[],
): Promise<void> {
  try {
    await this.client.upsert(this.collectionName, { points: points as any });
  } catch (error) {
    if (!this._hasBm25Slot) {
      throw error;
    }
    this._hasBm25Slot = false;
    console.warn(
      `Qdrant rejected the server-side BM25 vector for collection '${this.collectionName}'; ` +
        "disabling hybrid keyword search. This requires Qdrant >= 1.15.2 with inference enabled. " +
        "Retrying the write with a plain dense vector.",
    );
    const denseOnly = points.map((p) => ({
      ...p,
      vector: Array.isArray(p.vector) ? p.vector : p.vector[""],
    }));
    await this.client.upsert(this.collectionName, {
      points: denseOnly as any,
    });
  }
}
```
Sources: [mem0-ts/src/oss/src/vector_stores/qdrant.ts:320-343](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/qdrant.ts#L320-L343)

> [!NOTE]
> Qdrant client types predate server-side `sparse_vectors`. Because older server versions (Qdrant < 1.15.2 or unactivated clusters) reject server-side BM25 configs during upsert, the driver catches the first upsert failure, disables `_has_bm25_slot`, and transparently retries the write using a plain dense vector.

Sources: [mem0-ts/src/oss/src/vector_stores/qdrant.ts:312-343](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/qdrant.ts#L312-L343)

In Milvus, record construction for insertion and updates similarly checks `_has_bm25_schema`. If the schema supports BM25, the record includes a `text` field populated by extracting `text_lemmatized` or `data` from metadata and slicing it to a maximum of 65,535 characters.

```typescript
async insert(
  vectors: number[][],
  ids: string[],
  payloads: Record<string, any>[],
): Promise<void> {
  const data = vectors.map((vector, idx) => {
    const metadata = payloads[idx] || {};
    const row: Record<string, any> = {
      id: ids[idx],
      vectors: vector,
      metadata,
    };
    // Only write `text` when the collection has the BM25 schema; legacy
    // collections reject an unknown top-level field.
    if (this.hasBm25Schema) row.text = this.bm25Text(metadata);
    return row;
  });
  await this.client.insert({
    collection_name: this.collectionName,
    data,
  });
}
```
Sources: [mem0-ts/src/oss/src/vector_stores/milvus.ts:272-293](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/milvus.ts#L272-L293), [mem0/vector_stores/milvus.py:126-146](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/milvus.py#L126-L146)

## Pinecone Namespacing and Operations

### Index Resolution and Namespace Isolation

Pinecone client initialization resolves indices through either explicit client instances or API keys retrieved from parameters or the `PINECONE_API_KEY` environment variable. When creating or binding to collections, the driver checks existing indexes via `list_cols().names()` in Python, creating a new serverless or pod-based index if absent using either `ServerlessSpec` or `PodSpec`. Serverless configurations default to AWS region `us-west-2` if no explicit configuration dictionary is supplied. Namespace isolation is maintained by passing the configured namespace string directly into query, upsert, fetch, and delete operations.

Sources: [mem0/vector_stores/pinecone.py:58-123](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/pinecone.py#L58-L123)

> [!NOTE]
> Deletion behavior varies based on whether a namespace is configured. If a namespace is set, `delete_col()` clears only that specific namespace via `delete(delete_all=True, namespace=...)`. If `namespace` is `None`, it deletes the entire index collection.

Sources: [mem0/vector_stores/pinecone.py:361-371](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/pinecone.py#L361-L371), [mem0-ts/src/oss/src/vector_stores/pinecone.ts:318-330](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/pinecone.ts#L318-L330)

### Filter Operator Mapping and Migrations

Metadata filtering maps domain filter operators to Pinecone-compatible selector syntax. The driver translates comparison operators such as equality, bounds, and inclusion sets into standard Pinecone expression keys.

| Operator Key | Pinecone Selector |
| :--- | :--- |
| `eq` | `$eq` |
| `ne` | `$ne` |
| `gt` | `$gt` |
| `gte` | `$gte` |
| `lt` | `$lt` |
| `lte` | `$lte` |
| `in` | `$in` |
| `nin` | `$nin` |

Sources: [mem0/vector_stores/pinecone.py:189-222](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/pinecone.py#L189-L222), [mem0-ts/src/oss/src/vector_stores/pinecone.ts:209-242](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/pinecone.ts#L209-L242)

> [!WARNING]
> Text containment filters (`contains` and `icontains`) are unsupported by Pinecone metadata filters; the TypeScript driver skips them with a console warning, whereas unknown operators trigger an explicit runtime error.

Sources: [mem0-ts/src/oss/src/vector_stores/pinecone.ts:220-230](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/pinecone.ts#L220-L230)

## Weaviate Class and Filter Mechanics

### Weaviate Class Collections and Client Initialization

### Connection Schemes and Local Setup

The Weaviate driver supports three distinct connection pathways across both Python and TypeScript runtimes: local development instances, managed Weaviate Cloud instances via API keys, and custom HTTP/gRPC host targets. 

| Connection Mode | Trigger Condition | Initialization Method (Python) | Initialization Method (TypeScript) |
| :--- | :--- | :--- | :--- |
| **Local** | `cluster_url` contains `"localhost"` | `weaviate.connect_to_local()` | `weaviate.connectToLocal()` |
| **Cloud** | `auth_client_secret` or `apiKey` present | `weaviate.connect_to_weaviate_cloud()` | `weaviate.connectToWeaviateCloud()` |
| **Custom** | Default fallback when URL is provided | `weaviate.connect_to_custom()` | `weaviate.connectToCustom()` |

Sources: [mem0/vector_stores/weaviate.py:51-80](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/weaviate.py#L51-L80), [mem0-ts/src/oss/src/vector_stores/weaviate.ts:64-96](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/weaviate.ts#L64-L96)

> [!NOTE]
> When connecting via the custom fallback mode, the driver automatically targets gRPC port `50051` with insecure transport (`grpcSecure: false`), while inspecting the URL scheme to set the HTTP port to `443` for HTTPS or `8080` for plain HTTP.

Sources: [mem0/vector_stores/weaviate.py:60-79](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/weaviate.py#L60-L79), [mem0-ts/src/oss/src/vector_stores/weaviate.ts:81-95](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/weaviate.ts#L81-L95)

### Schema Properties and Collection Creation

Collections are configured with a standard set of text-typed properties and an HNSW vector index combined with a `none` vectorizer configuration.

```python
properties = [
    wvcc.Property(name="ids", data_type=wvcc.DataType.TEXT),
    wvcc.Property(name="hash", data_type=wvcc.DataType.TEXT),
    wvcc.Property(
        name="metadata",
        data_type=wvcc.DataType.TEXT,
        description="Additional metadata",
    ),
    wvcc.Property(name="data", data_type=wvcc.DataType.TEXT),
    wvcc.Property(name="created_at", data_type=wvcc.DataType.TEXT),
    wvcc.Property(name="category", data_type=wvcc.DataType.TEXT),
    wvcc.Property(name="updated_at", data_type=wvcc.DataType.TEXT),
    wvcc.Property(name="user_id", data_type=wvcc.DataType.TEXT),
    wvcc.Property(name="agent_id", data_type=wvcc.DataType.TEXT),
    wvcc.Property(name="run_id", data_type=wvcc.DataType.TEXT),
]

vectorizer_config = wvcc.Configure.Vectorizer.none()
vector_index_config = wvcc.Configure.VectorIndex.hnsw()

self.client.collections.create(
    self.collection_name,
    vectorizer_config=vectorizer_config,
    vector_index_config=vector_index_config,
    properties=properties,
)
```
Sources: [mem0/vector_stores/weaviate.py:131-156](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/weaviate.py#L131-L156), [mem0-ts/src/oss/src/vector_stores/weaviate.ts:18-29](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/weaviate.ts#L18-L29), [mem0-ts/src/oss/src/vector_stores/weaviate.ts:104-115](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/weaviate.ts#L104-L115)

### Vector Search Filter Compilation

Metadata search filters are compiled into Weaviate filter expressions by inspecting query keys for `user_id`, `agent_id`, and `run_id`. The drivers construct property-level equality checks and combine them using logical conjunction.

```python
collection = self.client.collections.get(str(self.collection_name))
filter_conditions = []
if filters:
    for key, value in filters.items():
        if value and key in ["user_id", "agent_id", "run_id"]:
            filter_conditions.append(Filter.by_property(key).equal(value))
combined_filter = Filter.all_of(filter_conditions) if filter_conditions else None
```
Sources: [mem0/vector_stores/weaviate.py:187-193](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/weaviate.py#L187-L193), [mem0-ts/src/oss/src/vector_stores/weaviate.ts:120-126](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/weaviate.ts#L120-L126)

> [!WARNING]
> During object insertion, the Python driver explicitly checks if an `ids` key exists within the payload dictionary and deletes it to prevent conflicts with Weaviate's native object UUID assignment.

Sources: [mem0/vector_stores/weaviate.py:175-178](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/weaviate.py#L175-L178)

## Turbopuffer and Upstash Serverless Models

### Serverless Namespace Queries and Row Parsing

Serverless vector backends delegate index lifecycle management and scaling to cloud infrastructure, exposing storage through lightweight namespace abstractions. The Turbopuffer driver structures vector collections as isolated namespaces that support implicit creation on first write. During similarity queries, result rows are parsed from model dumps where distance metrics—such as unbounded squared-L2 distances or cosine metrics—are normalized into bounded scores.

```python
results = []
for row in rows:
    row_dict = row.model_dump()
    row_id = str(row_dict.pop("id"))
    dist = row_dict.pop("$dist", None)
    row_dict.pop("vector", None)

    if dist is None:
        score = None
    elif self.distance_metric == "euclidean_squared":
        # $dist is unbounded squared-L2 (lower = closer); map to a bounded
        # higher-is-better score, mirroring milvus/baidu. Cosine returns 1 - dist.
        score = 1.0 / (1.0 + dist)
    else:
        score = 1 - dist

    results.append(OutputData(
        id=row_id,
        score=score,
        payload=row_dict,
    ))
return results
```
Sources: [mem0/vector_stores/turbopuffer.py:64-71](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/turbopuffer.py#L64-L71), [mem0/vector_stores/turbopuffer.py:118-139](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/turbopuffer.py#L118-L139), [mem0-ts/src/oss/src/vector_stores/turbopuffer.ts:68-80](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/turbopuffer.ts#L68-L80)

> [!NOTE]
> When the distance metric is configured to `euclidean_squared`, Turbopuffer returns unbounded squared-L2 values where lower numbers indicate closer proximity. The parser transforms these distances into higher-is-better relevance scores via the inverse mapping `1.0 / (1.0 + dist)`.

Sources: [mem0/vector_stores/turbopuffer.py:127-130](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/turbopuffer.py#L127-L130)

### Stringified Metadata Validation and Filter Construction

Upstash Vector enforces strict type validation for search and listing filters before transmitting expressions to the serverless endpoint. Filter keys are verified, and values must conform to primitive types (`str`, `int`, `float`, or `bool`). Furthermore, string values undergo strict validation to prevent query injection or syntax corruption.

```python
if not isinstance(value, (str, int, float, bool)):
    raise ValueError(
        f"Filter value for {key!r} must be str, int, float, or bool, "
        f"got {type(value).__name__}"
    )
if isinstance(value, str) and ('"' in value or "\\" in value):
    raise ValueError(
        f"Filter value for {key!r} contains prohibited characters "
        f"(double quote or backslash): {value!r}"
    )
```
Sources: [mem0/vector_stores/upstash_vector.py:23-32](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/upstash_vector.py#L23-L32), [mem0-ts/src/oss/src/vector_stores/upstash_vector.ts:229-239](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0-ts/src/oss/src/vector_stores/upstash_vector.ts#L229-L239)

> [!WARNING]
> Filter values of type `string` submitted to Upstash Vector cannot contain double quotes (`"`) or backslashes (`\`). Any matching characters trigger an explicit `ValueError` during filter preprocessing.

Sources: [mem0/vector_stores/upstash_vector.py:28-32](https://github.com/blade47/mem0/blob/b7ad69afda6b6ed030347c66d48a13e4de9dec08/mem0/vector_stores/upstash_vector.py#L28-L32)

## Related

- [Vector Store Architecture](https://www.doc0.dev/docs/faa36707-7c28-4f69-a18f-700ff61c704e/technical/vector-stores/vector-store-architecture)
- [Database Vector Stores](https://www.doc0.dev/docs/faa36707-7c28-4f69-a18f-700ff61c704e/technical/vector-stores/database-vector-stores)


## Sitemap

See the full [sitemap](https://www.doc0.dev/docs/faa36707-7c28-4f69-a18f-700ff61c704e/llms.txt) for all pages in this wiki.