Overseer Docs

Introduction

Overseer — a self-hosted AI assistant for your own server, explained in plain language.

Welcome to Overseer

Overseer is your own private AI assistant for a server or computer you control.

You can talk to it in a web chat, Telegram, or Discord. It can help you inspect files, answer questions about your system, and automate everyday tasks from one dashboard.

This documentation is written to help both:

  • people who just want to get it running without learning every technical detail
  • developers who want to customize, extend, or self-host it in production

Start Here

If you are not technical, this is the simplest way to think about it:

  • Overseer app: the actual AI assistant and dashboard
  • Docs site: the help website you are reading now

Use this guide based on what you want:

  • If you want to try Overseer itself, go to Quick Start
  • If you want to run Overseer for real on your own server, go to Deployment
  • If you want to publish only the docs on Vercel, go to Deployment

Why People Use Overseer

+It gives you one place to manage everything

Instead of jumping between terminals, config files, dashboards, and chat apps, you can manage your setup from one interface.

+You keep control

Overseer is self-hosted. Your settings, files, and providers stay under your control instead of living only inside a third-party hosted product.

+You can chat with it the way you like

Use the built-in web chat, Telegram, or Discord. You can also upload files, attach images, and work inside each user’s own file area.

+It can grow with you

Start with the basics, then add skills, more providers, bots, MCP servers, or custom workflows when you need them.

+It is built for real usage

It supports multiple users, separate sandboxes, permissions, audit logs, and production deployment patterns.

Quick Install

If you just want the fastest path:

curl -fsSL https://overseer.sh/install | bash

If you prefer doing it yourself:

git clone https://github.com/Quad-Labs-LLC/overseer.git
cd overseer
pnpm install
pnpm run db:init
cp .env.example .env
# Edit .env with your credentials, then:
pnpm dev

New here? Follow the Quick Start first. It is the shortest, least technical path.

On this page