OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
RAG & Knowledge Systems

Vectorwrap

Vectorwrap is an open-source Python library that provides a shared API for creating vector collections, inserting or updating embedding vectors with metadata, and querying similar vectors across supported database backends.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Vectorwrap — OpenViST project profile in RAG & Knowledge Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Vectorwrap when an application already produces embeddings and needs a common Python interface for storing and querying them across supported databases.

Who it is for

  • Python developers adding embedding retrieval to applications.
  • Teams building RAG systems that need vector search behind an existing or changeable database backend.
  • Developers prototyping vector search locally before moving to another supported database.

Documented uses

  • Store and update vectors with optional metadata.
  • Retrieve the nearest vectors by query embedding, with optional filtering.
  • Change a project's vector-search backend through a shared interface.
Before you choose

Requirements & limits

The repository is public and lists an MIT license. The documented installation uses pip.

  • PostgreSQL support requires pgvector; SQLite requires loadable extensions.
  • ClickHouse vector similarity indexes require version 25.8 or later and an experimental feature enabled.
  • Older MySQL versions use JSON arrays with Python-based distance calculations instead of native vector storage.
  • The application must provide an embedding function and vectors.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up a vector collection and query embeddings

Install Vectorwrap for a supported backend, then use its documented API to store and retrieve vectors.

Before you start

  • A Python environment and a supported database with a connection string.
  • An embedding function that returns vectors.
  • Backend-specific prerequisites: PostgreSQL requires pgvector; SQLite requires loadable extensions; DuckDB and ClickHouse use their respective Vectorwrap extras. ClickHouse vector similarity indexes require version 25.8 or later with the experimental feature enabled.

Steps

  1. Install vectorwrap for PostgreSQL or MySQL; add the documented sqlite, duckdb, or clickhouse extra for those backends.

  2. Import VectorDB and connect with the chosen backend’s connection string.

  3. Call create_collection with a collection name and vector dimension.

  4. Use upsert to store a vector and optional metadata, then call query with a query vector, result count and optional filter.

Expected outcome

A collection can hold embedding vectors and metadata, and a similarity query returns matching IDs and distances.

Guide limitations

  • Vectorwrap accepts vectors from an application-provided embedding function; it does not supply the embeddings.
  • Backend requirements and filtering behavior vary; SQLite filtering uses adaptive oversampling, while PostgreSQL and MySQL use native SQL filtering.
Documentation 1 ↗Documentation 2 ↗
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
mihirahuja1/vectorwrap · Official homepage title · retained in its source language
How we review sourcesSuggest a correction