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

Valori

Valori is an MIT-licensed Python vector database library for storing, indexing and searching high-dimensional vectors. Its documented components include multiple storage backends, indexing approaches, document parsers and a processing pipeline for cleaning, chunking and embedding content.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Valori — 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 Valori when a Python application needs a library-level component for embedding-based storage and retrieval, particularly when its documented parsing pipeline and choice of storage and index types fit the intended workflow.

Who it is for

  • Python developers adding vector storage and retrieval to an application.
  • Teams building semantic-search or document-retrieval workflows that need configurable storage and indexing.

Documented uses

  • Process text or PDF documents into embeddings and associate them with metadata.
  • Store vectors in memory, on disk or in a hybrid backend.
  • Index vectors and retrieve similar items from a vector query.
  • Choose among documented index types and distance metrics for a search workflow.
Before you choose

Requirements & limits

The repository provides pip and source-install instructions and states an MIT license.

  • Memory storage is non-persistent; disk storage is persistent and described as slower, while hybrid storage combines memory and disk.
  • Flat indexing is exhaustive and described as accurate but slower for large datasets; HNSW is described as approximate and fast for large datasets.
  • The README documents IVF, LSH and Annoy as additional index options, alongside scalar and product quantization.
  • The project’s README warns that it is in early development, with bugs and breaking changes expected.

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

Build a document-to-vector search workflow

Use the README’s example to process a PDF, store its embedding and search for similar documents.

Before you start

  • Python environment with Valori installed using the documented pip command
  • NumPy and the configured Sentence Transformers embedding model
  • A PDF document to process

Steps

  1. Install Valori with pip install valori.

  2. Create a MemoryStorage instance and a FlatIndex configured with the cosine metric, then initialize a VectorDBClient.

  3. Configure and initialize a ProcessingPipeline with parsing, cleaning, semantic chunking and the documented embedding model.

  4. Process the PDF, convert its embedding to a NumPy array, and insert it into the client with the document metadata.

  5. Process a text query, search using its embedding with k=5, then close the client and pipeline.

Expected outcome

The example returns up to five similar-document results with file names from their metadata.

Guide limitations

  • MemoryStorage is documented as non-persistent.
  • The README warns that Valori is in early development and that bugs and breaking changes are expected.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
varshith-Git/valori · Official homepage title · retained in its source language
How we review sourcesSuggest a correction