Artemis

Quick Start

Get the Artemis gateway and channels running locally.

Prerequisites

You need Node.js 22+ and an LLM API key. We support Gemini, OpenAI, Anthropic, or you can point it at a local Ollama instance if you have the hardware for it.

I strongly recommend using pnpm.

Installation

Clone the repo and install dependencies:

git clone https://github.com/eduardosilveiradev/artemis-ai
cd artemis-ai
pnpm install
pnpm setup:cli

Copy the environment template:

cp .env.example .env

Open .env and configure your API keys. You only need to set the ones you actually plan to use:

GOOGLE_API_KEY=your-gemini-key
OPENAI_API_KEY=your-openai-key
OLLAMA_BASE_URL=http://localhost:11434

Starting the Gateway

Everything routes through the Artemis gateway. When you run this, it boots up the core agent, the MCP servers, and the scheduler daemon.

pnpm dev

The Web UI comes up at http://localhost:4080. The underlying HTTP/WebSocket gateway binds to 4081.

Connecting Channels

Once the gateway is up, you need a way to talk to it. Pick whichever interface you prefer:

InterfaceCommandNotes
Web UIpnpm webRuns on port 4080
CLIartemis chatNative interactive terminal layout
Discordpnpm discordDM-only bot
Telegrampnpm telegramBot; requires auth on first run
WhatsApppnpm whatsappRequires QR scan on first run
Gmailpnpm gmailRequires OAuth2 consent

For the CLI and Web UI, you don't need any special configuration. The other channels require their respective tokens or OAuth setups in your .env.

I usually keep the gateway running in a background process, and then spin up the CLI when I need something quick, or leave the Discord bot running for mobile access.