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

Lodedb

LodeDB is an embedded vector database that stores embeddings on disk and lets applications add, retrieve, and search them locally, including through text, metadata filters, hybrid retrieval, and multimodal search.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Lodedb — 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 LodeDB when an application needs an embedded, persistent vector store for local retrieval or AI memory, especially when on-disk storage, metadata filtering, hybrid search, or multimodal retrieval matters.

Who it is for

  • Developers building local retrieval-augmented generation (RAG) applications.
  • Teams adding durable vector memory to LangChain, LlamaIndex, or mem0 projects.
  • Developers who need local text, image, or hybrid embedding search.

Documented uses

  • Persist embeddings and search them from an application process.
  • Retrieve relevant text for local RAG workflows.
  • Filter vector-search results using metadata conditions.
  • Combine vector recall with exact lexical matches for identifiers and other terms.
  • Use a shared CLIP space to search images and text together.
Before you choose

Requirements & limits

The repository identifies the core as Apache-2.0 licensed and documents Python 3.11+ with prebuilt wheels for Linux, macOS, and Windows.

  • The database is embedded and in-process, with on-disk persistence; the documented setup does not require a separate database daemon.
  • Built-in text embeddings are available through the embeddings extra; the base installation is intended for users supplying their own vectors or embedding model.
  • GPU-resident batched search is documented for Linux with CUDA, while embedding acceleration depends on the selected runtime and installed packages.
  • The repository describes exact-by-default retrieval, but its benchmark results are reported for specific workloads and hardware; assess performance against your own data and environment.
  • The README documents framework adapters for LangChain, LlamaIndex, and mem0, plus a CLI and MCP server.

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

Install LodeDB and confirm the installation

Install the Python package with its built-in text embedding support and run the documented installation check.

Before you start

  • Python 3.11 or later.
  • A Linux, macOS, or Windows environment supported by the prebuilt wheels.

Steps

  1. Install with the command below.

    Documented command
    pip install "lodedb[embeddings]"

  2. Run lodedb doctor to confirm the installation.

Expected outcome

The package is installed with built-in text embedding support, and the documented doctor command provides an installation check.

Guide limitations

  • This guide covers installation and the check only; it does not configure or run a retrieval workflow.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Vector search

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
LodedbPersist embeddings and search them from an application process.Linux, macOS, Self-hosted, ServerThe repository identifies the core as Apache-2.0 licensed and documents Python 3.11+ with prebuilt wheels for Linux, macOS, and Windows.
MettisIndex text documents through an HTTP endpoint.Linux, Self-hosted, Server
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
Egoist-Machines/LodeDB · Official homepage title · retained in its source language
How we review sourcesSuggest a correction