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

AI Memory

AI Memory is a self-hosted memory server and framework for AI-agent applications. It processes recent conversation context through a short-term, staging and long-term funnel, using LLM scoring, recurrence and time-window checks to filter candidate memories, with semantic search over long-term records.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

AI Memory — 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 it when an agent application needs to carry selected information across sessions and retrieve it later, with a staged process for filtering, promotion and decay rather than indiscriminate long-term retention.

Who it is for

  • Developers adding persistent memory to an AI-agent application.
  • Teams designing a staged memory pipeline with configurable filtering and retention.
  • Builders using retrieval-augmented generation with a Qdrant-backed memory store.

Documented uses

  • Keep recent conversation context in a short-term store.
  • Evaluate candidate memories before promoting them to long-term storage.
  • Retrieve semantically relevant stored memories for later agent interactions.
  • Monitor memory-system metrics, trends and service health through the dashboard.
Before you choose

Requirements & limits

The repository is public and its metadata specifies an Apache-2.0 license. The documented deployment is self-hosted.

  • The documented stack requires Go 1.25+, Redis 7.0+, MySQL 8.0+ and Qdrant 1.0+.
  • An OpenAI API key or compatible endpoint is listed as a prerequisite for LLM-based judgment.
  • The README describes Redis for short-term and staging data, Qdrant for semantic long-term storage, and MySQL for metadata and system state.
  • Retention and promotion are configurable, including short-term expiry and long-term decay settings.

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 and start AI Memory locally

Prepare the repository, configure its services and build and start the memory server.

Before you start

  • Go 1.25+
  • Redis 7.0+
  • MySQL 8.0+
  • Qdrant 1.0+
  • An OpenAI API key or compatible endpoint

Steps

  1. Clone the repository and change into its directory.

  2. Copy .env.example to .env and edit it with API keys and database credentials.

  3. Apply the database schema with mysql -u root -p < schema.sql.

  4. Run the command below, then build with the command below.

    Documented command
    go mod download
    Documented command
    go build -o ai-memory

  5. Start the server with ./ai-memory.

Expected outcome

The README says the server starts at `http://localhost:8080`.

Guide limitations

  • This is a setup recipe; it does not verify that the service runs in a particular environment.
  • The database and LLM credentials must be configured for the deployment.
Documentation 1 ↗
Source-based practical guide

Start the application stack with Docker Compose

Configure the project and start the application with Redis, MySQL and Qdrant through the documented Compose setup.

Before you start

  • Docker Compose
  • An OpenAI API key or compatible endpoint configured in .env

Steps

  1. Clone the repository and change into its directory.

  2. Copy .env.example to .env and configure the API key and other required environment values.

  3. Change to the docker directory and run docker-compose up -d.

  4. Use docker-compose logs -f app to view application logs.

Expected outcome

The README describes the application and its Redis, MySQL and Qdrant services starting together, with the application on port 8080.

Guide limitations

  • This recipe follows the repository's documented Compose setup and does not establish compatibility with other environments.
  • The README's instructions require configuring environment values before starting the stack.
Documentation 1 ↗
Evidence and transparency

Sources & review

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