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

Memary

Memary is an open-source Python library for adding memory to autonomous agents. Its ChatAgent combines persona files, past-chat history, a memory stream and an entity knowledge store, with a documented architecture that includes a routing agent and knowledge graph.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Memary — 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 Memary when building an agent application that needs memory and knowledge context configured in code, including a documented multi-agent pattern using FalkorDB.

Who it is for

  • Developers building autonomous-agent applications that need stored conversation or entity context.
  • Teams exploring knowledge-graph-backed memory or separate memory contexts for multiple agents.
  • Python developers who want to configure model choices and built-in or custom agent tools.

Documented uses

  • Initialize an agent with persona, chat-history, memory-stream and entity-knowledge files.
  • Select from built-in search, vision, location and stocks tools, or add and remove custom tools.
  • Configure local Ollama models or the listed GPT model alternatives.
  • Set up multiple agent graphs with FalkorDB and distinct user IDs.
Before you choose

Requirements & limits

The repository is public and identifies the project as open source under the MIT license.

  • The README specifies Python 3.11.9 or earlier.
  • Ollama-hosted local models are the default unless explicitly specified otherwise; the README also lists GPT-3.5 Turbo and GPT-4 Vision Preview.
  • The README documents FalkorDB for multi-graph agent setups and lists both FalkorDB and Neo4j database connection settings.
  • The included app is launched through Streamlit; the examples also show direct ChatAgent initialization in Python.
  • Configuration can involve credentials for external services, depending on the integrations used.

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 Memary and configure the included app

Prepare the documented Python project environment, choose model and database options, and start its Streamlit app.

Before you start

  • Python 3.11.9 or earlier
  • For local model use, Ollama and a supported local model; the README also lists GPT-3.5 Turbo and GPT-4 Vision Preview alternatives
  • A .env file containing credentials needed for the selected integrations
  • A FalkorDB or Neo4j connection when using a graph database

Steps

  1. Install with pip using the command below, or create a virtual environment and install dependencies with the command below.

    Documented command
    pip install memary
    Documented command
    pip install -r requirements.txt

  2. Select the documented language and vision model options; local Ollama models are the default unless explicitly specified otherwise.

  3. Set the needed credentials and database connection values in .env, then update streamlit_app/data/user_persona.txt from the provided template. Optionally edit the system persona.

  4. Run cd streamlit_app followed by streamlit run app.py.

Expected outcome

The README’s Streamlit app is launched with the configured model, credentials and persona data.

Guide limitations

  • The README lists credentials for several integrations but does not say that all are required for every setup.
  • The instructions do not document deployment beyond running the Streamlit app.
  • The guide reflects repository instructions and is not a human execution test.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
kingjulio8238/Memary · Official homepage title · retained in its source language
How we review sourcesSuggest a correction