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

Hypervectordb

Hypervectordb is a C# library for storing vectors alongside associated documents in a local database and retrieving top matches with similarity or distance queries. Applications can index strings or line-based text files, organize content into named indices, and save or reload the database from a folder path.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Hypervectordb — 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 it when a C# application needs local vector-backed document retrieval, including as a retrieval layer in a larger RAG workflow. It fits projects where the application can provide an Embedder and manage the surrounding indexing and generation logic.

Who it is for

  • C# developers building local document retrieval into an application.
  • Developers assembling RAG systems who need a vector-search component and will provide the surrounding generation workflow.
  • Teams that want to index text files line by line and retain file-path and line-number references in indexed records.

Documented uses

  • Store text with its embedding vectors in a local database.
  • Index individual strings or lines from text files, with optional preprocessing and postprocessing.
  • Query indexed content and retrieve a configurable number of matching documents with scores.
  • Separate indexed content across named indices and save or reload database data.
Before you choose

Requirements & limits

The repository is public and identifies the project as MIT-licensed. The README notes that some code in Math.cs is based on Acord.Math, which is licensed under LGPL v2.1.

  • The library requires an Embedder object and a database folder path; the README’s example uses an LM Studio embedder.
  • Text-file indexing splits files into lines and can use custom preprocessing and postprocessing methods.
  • The query methods return top-k documents and scores; the documented default for k is 5.
  • Query caching is described only for cosine similarity queries.
  • Automatic parallelization is described as configurable; disk saves use LZ4 compression. The README labels its features section “To be updated.”

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 local index and query text

Add the library to a .NET project and follow the README’s core flow for indexing and querying text.

Before you start

  • A .NET project
  • An Embedder object
  • A folder path for the database

Steps

  1. Add the package with dotnet add package HyperVectorDB.

  2. Initialize HyperVectorDB with an Embedder object and a folder path, then create a named index.

  3. Index a string or text file; the file method indexes lines and associates them with the file path and line number.

  4. Save or load the database as needed, then call a query method with a text string and maximum result count.

Expected outcome

A local index of text that can return matching documents and their similarity or distance values.

Guide limitations

  • The guide describes the library’s documented usage flow, not a tested execution.
  • An Embedder is required; the README example uses Embedder.LmStudio().
  • Cosine, Jaccard, Euclidean, Manhattan, Chebyshev and Canberra are the documented query measures.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
17 October 2026
Original website title
deatos/HyperVectorDB · Official homepage title · retained in its source language
How we review sourcesSuggest a correction