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

Stash

Stash is a self-hosted persistent-memory service for AI agents. It stores episodes and facts in PostgreSQL with pgvector, exposes memory operations through an MCP server, and runs background consolidation that turns observations into structured knowledge.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Stash — 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 Stash when an MCP-compatible agent needs persistent, queryable memory and you want a service that can consolidate new observations into structured knowledge. The self-hosted route fits teams prepared to configure models and run Docker Compose with PostgreSQL and pgvector.

Who it is for

  • Builders of MCP-compatible AI agents that need memory to persist across sessions.
  • Operators seeking a self-hosted memory service backed by PostgreSQL and pgvector.
  • Developers configuring local embeddings and reasoning through the documented Ollama setup.

Documented uses

  • Give an agent a way to remember and recall information across sessions.
  • Store episodes and facts in a PostgreSQL-backed vector memory store.
  • Consolidate observations into facts, relationships, patterns and other structured knowledge.
  • Connect an MCP-compatible client to memory operations over SSE.
Before you choose

Requirements & limits

The repository is public and licensed under Apache 2.0. Its README also describes a separate hosted version at usestash.io as in beta and free during beta; that service is written from scratch and its features differ from the repository version.

  • The self-hosted setup requires configuring a model and running Docker Compose.
  • The MCP server is documented over SSE at http://localhost:8080/sse.
  • The README provides an Ollama guide for a local embeddings-and-reasoning configuration.
  • The repository's hosted counterpart is a separate implementation with a different feature set.

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

Set up Stash with Docker Compose

Start the self-hosted Stash service and its documented PostgreSQL, pgvector and MCP components.

Before you start

  • Git
  • Docker Compose
  • An API key and model for the configured service, unless following the separate local Ollama setup

Steps

  1. Clone the Stash repository and change into its directory.

  2. Copy .env.example to .env and edit it with the API key and model.

  3. Run the command below.

    Documented command
    docker compose up

Expected outcome

The README says this starts PostgreSQL with pgvector, runs migrations, and brings up the MCP server with background consolidation. Continue to the Getting Started guide to connect an MCP client and use init, remember and recall.

Guide limitations

  • These steps cover starting the service, not completing or verifying an agent integration.
  • Ollama-based local configuration uses a separate setup guide.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
alash3al/stash · Official homepage title · retained in its source language
How we review sourcesSuggest a correction