OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Memory & RAG

Vektori

Vektori is an open-source Python library for AI-agent memory. It organizes recollections in a three-layer sentence graph linking concise facts, episodes and raw source sentences, and supports adding and searching conversation messages. The project also describes file-system memory.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Vektori — OpenViST project profile in Agent Memory & RAG.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Vektori when an agent needs to recall details across sessions while retaining links to the conversations those details came from.

Who it is for

  • Developers building AI agents that need persistent, user-specific memory.
  • Teams that want to retrieve conversational context alongside summarized facts.

Documented uses

  • Store conversation messages with user and session identifiers.
  • Search a user's memory for relevant facts and episodes.
  • Choose retrieval depth, from facts-only lookups to answers with source sentences and surrounding context.
Before you choose

Requirements & limits

The repository identifies Vektori as open source under the Apache-2.0 license. The README specifies Python 3.10 or later and documents installation through pip.

  • SQLite is the default database; the README lists optional Postgres, Neo4j, Qdrant and Milvus support.
  • The documented quickstart configures OpenAI embedding and extraction models.
  • The project describes file-system memory, while the quickstart shows conversation-message memory operations.

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

Add conversation memory and retrieve it

Store messages for a user and search the resulting memory for relevant facts and episodes.

Before you start

  • Python 3.10 or later.
  • Install Vektori with pip install vektori.
  • The quickstart configures OpenAI embedding and extraction models.

Steps

  1. Create a Vektori instance with an embedding model and extraction model.

  2. Pass conversation messages to add with a session ID and user ID.

  3. Call search with a query, the user ID and a retrieval depth such as l1.

  4. Read the returned facts and episodes, then close the Vektori instance.

Expected outcome

The example retrieves concise facts and an episode from the stored conversation, including the source context represented by the selected retrieval depth.

Guide limitations

  • The example relies on the named OpenAI models; other model configurations are not detailed in this quickstart.
  • The example demonstrates adding and searching conversation messages, not file-system memory setup.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
12 October 2026
Original website title
bandr-ai/vektori · Official homepage title · retained in its source language
How we review sourcesSuggest a correction