Self-Hosting

Overview

MemLib is fully open-source and can be self-hosted. This guide covers deploying the complete stack — the Next.js web dashboard, API routes, and database setup.

Prerequisites

Node.js 18+
PostgreSQL with pgvector extension
An AI provider API key (Gemini, OpenAI, etc.)

Step 1: Clone and Install

git clone https://github.com/yourusername/memlib.git
cd memlib
npm install

Step 2: Configure Environment

Copy the example env file and fill in your values:

cp apps/web/.env.example apps/web/.env

Required environment variables:

# Database (your central MemLib database — stores projects, API keys, users)
DATABASE_URL=postgres://user:pass@host:5432/memlib

# Auth
BETTER_AUTH_SECRET=your-secret-key
BETTER_AUTH_URL=http://localhost:8001

# OAuth providers (at least one)
GOOGLE_CLIENT_ID=...
GOOGLE_CLIENT_SECRET=...
GITHUB_CLIENT_ID=...
GITHUB_CLIENT_SECRET=...

Step 3: Push Database Schema

npm run db:push

This creates the project, API key, and user tables in your central database.

Step 4: Run in Development

npm run dev

Dashboard: http://localhost:8001
API: http://localhost:8001/api

Step 5: Deploy to Production

Vercel (Recommended)

The Next.js app deploys directly to Vercel:

Push to GitHub
Import the repo in Vercel
Set the root directory to apps/web
Add environment variables
Deploy

Docker

FROM node:20-alpine
WORKDIR /app
COPY . .
RUN npm install
RUN npm run build
EXPOSE 3000
CMD ["npm", "start"]

Architecture Notes

Central database — stores projects, users, API keys, and auth data
Per-project databases — each project connects to its own PostgreSQL instance (BYOD model). Users provide their own database URL when creating a project.
AI providers — configured per-project. Users bring their own API keys for embedding and LLM providers.

This means your self-hosted MemLib instance is a control plane — it manages projects and routes requests, but memory data lives in each user's own database.

On this page